id
string | submitter
string | authors
string | title
string | comments
string | journal-ref
string | doi
string | report-no
string | categories
string | license
string | abstract
string | versions
list | update_date
timestamp[s] | authors_parsed
sequence |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2506.09176 | Haoyuan Cai | Haoyuan Cai, Zhenghao Peng, Bolei Zhou | Robot-Gated Interactive Imitation Learning with Adaptive Intervention Mechanism | ICML 2025 Poster | null | null | null | cs.AI cs.LG cs.RO | http://creativecommons.org/licenses/by/4.0/ | Interactive Imitation Learning (IIL) allows agents to acquire desired behaviors through human interventions, but current methods impose high cognitive demands on human supervisors. We propose the Adaptive Intervention Mechanism (AIM), a novel robot-gated IIL algorithm that learns an adaptive criterion for requesting human demonstrations. AIM utilizes a proxy Q-function to mimic the human intervention rule and adjusts intervention requests based on the alignment between agent and human actions. By assigning high Q-values when the agent deviates from the expert and decreasing these values as the agent becomes proficient, the proxy Q-function enables the agent to assess the real-time alignment with the expert and request assistance when needed. Our expert-in-the-loop experiments reveal that AIM significantly reduces expert monitoring efforts in both continuous and discrete control tasks. Compared to the uncertainty-based baseline Thrifty-DAgger, our method achieves a 40% improvement in terms of human take-over cost and learning efficiency. Furthermore, AIM effectively identifies safety-critical states for expert assistance, thereby collecting higher-quality expert demonstrations and reducing overall expert data and environment interactions needed. Code and demo video are available at https://github.com/metadriverse/AIM. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 18:43:26 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Cai",
"Haoyuan",
""
],
[
"Peng",
"Zhenghao",
""
],
[
"Zhou",
"Bolei",
""
]
] |
2506.09177 | Kansei Inamura | Kansei Inamura, Sheng-Jie Huang, Apoorv Tiwari, Sakura Schafer-Nameki | (2+1)d Lattice Models and Tensor Networks for Gapped Phases with Categorical Symmetry | 58 pages self-contained summary + 100 pages main text + appendices | null | null | null | cond-mat.str-el hep-th math-ph math.MP math.QA quant-ph | http://creativecommons.org/licenses/by/4.0/ | Gapped phases in 2+1 dimensional quantum field theories with fusion 2-categorical symmetries were recently classified and characterized using the Symmetry Topological Field Theory (SymTFT) approach arXiv:2408.05266, arXiv:2502.20440. In this paper, we provide a systematic lattice model construction for all such gapped phases. Specifically, we consider ``All boson type" fusion 2-category symmetries, all of which are obtainable from 0-form symmetry groups $G$ (possibly with an 't Hooft anomaly) via generalized gauging--that is, by stacking with an $H$-symmetric TFT and gauging a subgroup $H$. The continuum classification directly informs the lattice data, such as the generalized gauging that determines the symmetry category, and the data that specifies the gapped phase. We construct commuting projector Hamiltonians and ground states applicable to any non-chiral gapped phase with such symmetries. We also describe the ground states in terms of tensor networks. In light of the length of the paper, we include a self-contained summary section presenting the main results and examples. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 18:44:46 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Inamura",
"Kansei",
""
],
[
"Huang",
"Sheng-Jie",
""
],
[
"Tiwari",
"Apoorv",
""
],
[
"Schafer-Nameki",
"Sakura",
""
]
] |
2506.09178 | Denis Zhidkikh | Denis Zhidkikh, Ville Isom\"ott\"onen, Toni Taipalus | Understanding Self-Regulated Learning Behavior Among High and Low Dropout Risk Students During CS1: Combining Trace Logs, Dropout Prediction and Self-Reports | 29 pages, 8 figures, 3 tables, submitted to ACM Transactions on Computing Education | null | null | null | cs.CY | http://creativecommons.org/licenses/by/4.0/ | The introductory programming course (CS1) at the university level is often perceived as particularly challenging, contributing to high dropout rates among Computer Science students. Identifying when and how students encounter difficulties in this course is critical for providing targeted support. This study explores the behavioral patterns of CS1 students at varying dropout risks using self-regulated learning (SRL) as the theoretical framework. Using learning analytics, we analyzed trace logs and task performance data from a virtual learning environment to map resource usage patterns and used student dropout prediction to distinguish between low and high dropout risk behaviors. Data from 47 consenting students were used to carry out the analysis. Additionally, self-report questionnaires from 29 participants enriched the interpretation of observed patterns. The findings reveal distinct weekly learning strategy types and categorize course behavior. Among low dropout risk students, three learning strategies were identified that different in how students prioritized completing tasks and reading course materials. High dropout risk students exhibited nine different strategies, some representing temporary unsuccessful strategies that can be recovered from, while others indicating behaviors of students on the verge of dropping out. This study highlights the value of combining student behavior profiling with predictive learning analytics to explain dropout predictions and devise targeted interventions. Practical findings of the study can in turn be used to help teachers, teaching assistants and other practitioners to better recognize and address students at the verge of dropping out. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 18:46:45 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Zhidkikh",
"Denis",
""
],
[
"Isomöttönen",
"Ville",
""
],
[
"Taipalus",
"Toni",
""
]
] |
2506.09179 | Evgenii Karkaryan | E. K. Karkaryan, V. F. Obraztsov, M. I. Vysotsky | Does the existence of $f_0$ meson enlarge the $K^+ \to \pi^0\pi^0\pi^0 e^+ \nu$ decay probability? | 3 pages, 1 figure, 1 table | null | null | null | hep-ph hep-ex | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We estimate a possible contribution of the two-pion system $f_0(500)$ wide resonance to the decay probability of the super rare decay $K^+ \to 3\pi^0 e^+ \nu$. We vary the width of the resonance from 0.01 MeV to 1000 MeV. We found that the contribution of the narrow resonance in a system of several final state particles into the decay probability is proportional to its width. However it never exceeds the value of the decay probability calculated on free final state particles neglecting the resonance contribution. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 18:50:40 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Karkaryan",
"E. K.",
""
],
[
"Obraztsov",
"V. F.",
""
],
[
"Vysotsky",
"M. I.",
""
]
] |
2506.09180 | Khai Doan | Khai Doan, Wesley Araujo, Evangelos Kranakis, Ioannis Lambadaris, Yannis Viniotis, Wonjae Shin | Optimal Task Offloading with Firm Deadlines for Mobile Edge Computing Systems | null | null | null | null | eess.SY cs.SY | http://creativecommons.org/licenses/by/4.0/ | Under a dramatic increase in mobile data traffic, a promising solution for edge computing systems to maintain their local service is the task migration that may be implemented by means of Autonomous mobile agents (AMA). In designing an optimal scheme for task offloading to AMA, we define a system cost as a minimization objective function that comprises two parts. First, an offloading cost which can be interpreted as the cost of using computational resources from the AMA. Second, a penalty cost due to potential task expiration. To minimize the expected (timeaverage) cost over a given time horizon, we formulate a Dynamic programming (DP). However, the DP Equation suffers from the well-known curse of dimensionality, which makes computations intractable, especially for infinite system state space. To reduce the computational burden, we identify three important properties of the optimal policy and show that it suffices to evaluate the DP Equation on a finite subset of the state space only. We then prove that the optimal task offloading decision at a state can be inferred from that at its adjacent states, further reducing the computational load. We present simulations to verify the theoretical results and to provide insights into the considered system. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 18:55:00 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Doan",
"Khai",
""
],
[
"Araujo",
"Wesley",
""
],
[
"Kranakis",
"Evangelos",
""
],
[
"Lambadaris",
"Ioannis",
""
],
[
"Viniotis",
"Yannis",
""
],
[
"Shin",
"Wonjae",
""
]
] |
2506.09181 | Pablo Ramirez-Espinosa | Pablo Ram\'irez-Espinosa, David Morales-Jim\'enez, Beatriz Soret | Energy efficiency of DMAs vs. conventional MIMO: a sensitivity analysis | null | null | null | null | eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Motivated by the stringent and challenging need for `greener communications' in increasingly power-hungry 5G networks, this paper presents a detailed energy efficiency analysis for three different multi-antenna architectures, namely fully-digital arrays, hybrid arrays, and dynamic metasurface antennas (DMAs). By leveraging a circuital model, which captures mutual coupling, insertion losses, propagation through the waveguides in DMAs and other electromagnetic phenomena, we design a transmit Wiener filter solution for the three systems. We then use these results to analyze the energy efficiency, considering different consumption models and supplied power, and with particular focus on the impact of the physical phenomena. DMAs emerge as an efficient alternative to classical arrays across diverse tested scenarios, most notably under low transmission power, strong coupling, and scalability requirements. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 18:55:02 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Ramírez-Espinosa",
"Pablo",
""
],
[
"Morales-Jiménez",
"David",
""
],
[
"Soret",
"Beatriz",
""
]
] |
2506.09182 | Hang Zhou | Hang Zhou, Chengyuan Ma, Shiyu Shen, Xiaopeng Li | Towards Full-Scenario Safety Evaluation of Automated Vehicles: A Volume-Based Method | NA | null | null | null | cs.RO cs.ET | http://creativecommons.org/licenses/by/4.0/ | With the rapid development of automated vehicles (AVs) in recent years, commercially available AVs are increasingly demonstrating high-level automation capabilities. However, most existing AV safety evaluation methods are primarily designed for simple maneuvers such as car-following and lane-changing. While suitable for basic tests, these methods are insufficient for assessing high-level automation functions deployed in more complex environments. First, these methods typically use crash rate as the evaluation metric, whose accuracy heavily depends on the quality and completeness of naturalistic driving environment data used to estimate scenario probabilities. Such data is often difficult and expensive to collect. Second, when applied to diverse scenarios, these methods suffer from the curse of dimensionality, making large-scale evaluation computationally intractable. To address these challenges, this paper proposes a novel framework for full-scenario AV safety evaluation. A unified model is first introduced to standardize the representation of diverse driving scenarios. This modeling approach constrains the dimension of most scenarios to a regular highway setting with three lanes and six surrounding background vehicles, significantly reducing dimensionality. To further avoid the limitations of probability-based method, we propose a volume-based evaluation method that quantifies the proportion of risky scenarios within the entire scenario space. For car-following scenarios, we prove that the set of safe scenarios is convex under specific settings, enabling exact volume computation. Experimental results validate the effectiveness of the proposed volume-based method using both AV behavior models from existing literature and six production AV models calibrated from field-test trajectory data in the Ultra-AV dataset. Code and data will be made publicly available upon acceptance of this paper. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 18:58:08 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Zhou",
"Hang",
""
],
[
"Ma",
"Chengyuan",
""
],
[
"Shen",
"Shiyu",
""
],
[
"Li",
"Xiaopeng",
""
]
] |
2506.09183 | Devin White | Mingkang Wu, Devin White, Evelyn Rose, Vernon Lawhern, Nicholas R Waytowich and Yongcan Cao | Multi-Task Reward Learning from Human Ratings | Accepted to the workshop on Models of Human Feedback for AI Alignment at the 42nd International Conference on Machine Learning | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Reinforcement learning from human feeback (RLHF) has become a key factor in aligning model behavior with users' goals. However, while humans integrate multiple strategies when making decisions, current RLHF approaches often simplify this process by modeling human reasoning through isolated tasks such as classification or regression. In this paper, we propose a novel reinforcement learning (RL) method that mimics human decision-making by jointly considering multiple tasks. Specifically, we leverage human ratings in reward-free environments to infer a reward function, introducing learnable weights that balance the contributions of both classification and regression models. This design captures the inherent uncertainty in human decision-making and allows the model to adaptively emphasize different strategies. We conduct several experiments using synthetic human ratings to validate the effectiveness of the proposed approach. Results show that our method consistently outperforms existing rating-based RL methods, and in some cases, even surpasses traditional RL approaches. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:00:19 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Wu",
"Mingkang",
""
],
[
"White",
"Devin",
""
],
[
"Rose",
"Evelyn",
""
],
[
"Lawhern",
"Vernon",
""
],
[
"Waytowich",
"Nicholas R",
""
],
[
"Cao",
"Yongcan",
""
]
] |
2506.09184 | Aklant Bhowmick | Aklant K. Bhowmick, Laura Blecha, Luke Z. Kelley, Aneesh Sivasankaran, Paul Torrey, Rainer Weinberger, Nianyi Chen, Mark Vogelsberger, Lars Hernquist, and Priyamvada Natarajan | Dynamics of low-mass black hole seeds in the BRAHMA simulations using subgrid-dynamical friction: Impact on merger-driven black hole growth in the high redshift Universe | 30 pages, 17 figures | null | null | null | astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | We analyze the dynamics of low-mass black hole (BH) seeds in the high-redshift ($z\gtrsim5$) Universe using a suite of $[4.5~\mathrm{Mpc}]^3$ and $[9~\mathrm{Mpc}]^3$ BRAHMA cosmological hydrodynamic simulations. The simulations form seeds with mass $M_{\mathrm{seed}}=2.2\times10^3~M_{\odot}$ in halos that exceed critical thresholds of dense & metal-poor gas mass ($5-150~M_{\mathrm{seed}}$) and the halo mass ($1000-10000~M_{\mathrm{seed}}$). While the initial BRAHMA boxes pinned the BHs to the halo centers, here we implement a sub-grid dynamical friction (DF) model. We also compare simulations where the BH is allowed to wander without the added DF. We investigate the spatial and velocity offsets of BHs in their host subhalos, as well as BH merger rates. We find that subgrid DF is crucial to ensure that a significant fraction of BHs effectively sink to halo centers by $z\sim5$, thereby enabling them to get gravitationally bound and merge with other BHs at separations close to the spatial resolution ($\sim0.2-0.4~\rm kpc$) of the simulation. For the BHs that merge, the associated merger time scales lag between $\sim100-1000~\mathrm{Myr}$ after their host halos merge. Compared to predictions using BH repositioning, the overall $z\gtrsim5$ BH merger rates under subgrid DF decrease by a factor of $\sim4-10$. Under subgrid DF, the different seed models predict merger rates between $\sim100-1000$ events per year at $z\gtrsim5$. These mergers dominate early BH growth, assembling BHs up to $\sim10^4-10^5~M_{\odot}$ by $z\sim5$, wherein $\lesssim2~\%$ of their mass is assembled via gas accretion. Our results highlight the promise for constraining seeding mechanisms using gravitational waves from future facilities such as the Laser Interferometer Space Antenna. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:01:42 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Bhowmick",
"Aklant K.",
""
],
[
"Blecha",
"Laura",
""
],
[
"Kelley",
"Luke Z.",
""
],
[
"Sivasankaran",
"Aneesh",
""
],
[
"Torrey",
"Paul",
""
],
[
"Weinberger",
"Rainer",
""
],
[
"Chen",
"Nianyi",
""
],
[
"Vogelsberger",
"Mark",
""
],
[
"Hernquist",
"Lars",
""
],
[
"Natarajan",
"Priyamvada",
""
]
] |
2506.09185 | Sreyoshi Bhaduri | Rubaina Khan, Tammy Mackenzie, Sreyoshi Bhaduri, Animesh Paul, Branislav Radelji\'c, Joshua Owusu Ansah, Beyza Nur Guler, Indrani Bhaduri, Rodney Kimbangu, Nils Ever Murrugarra Llerena, Hayoung Shin, Lilianny Virguez, Rosa Paccotacya Yanque, Thomas Mekha\"el, Allen Munoriyarwa, Leslie Salgado, Debarati Basu, Curwyn Mapaling, Natalie Perez, Yves Gaudet, Paula Larrondo | Whole-Person Education for AI Engineers | conference pre-print, position paper. 21 pages | null | null | null | cs.CY | http://creativecommons.org/licenses/by/4.0/ | This autoethnographic study explores the need for interdisciplinary education spanning both technical and philosophical skills - as such, this study leverages whole-person education as a theoretical approach needed in AI engineering education to address the limitations of current paradigms that prioritize technical expertise over ethical and societal considerations. Drawing on a collaborative autoethnography approach of fourteen diverse stakeholders, the study identifies key motivations driving the call for change, including the need for global perspectives, bridging the gap between academia and industry, integrating ethics and societal impact, and fostering interdisciplinary collaboration. The findings challenge the myths of technological neutrality and technosaviourism, advocating for a future where AI engineers are equipped not only with technical skills but also with the ethical awareness, social responsibility, and interdisciplinary understanding necessary to navigate the complex challenges of AI development. The study provides valuable insights and recommendations for transforming AI engineering education to ensure the responsible development of AI technologies. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:03:31 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Khan",
"Rubaina",
""
],
[
"Mackenzie",
"Tammy",
""
],
[
"Bhaduri",
"Sreyoshi",
""
],
[
"Paul",
"Animesh",
""
],
[
"Radeljić",
"Branislav",
""
],
[
"Ansah",
"Joshua Owusu",
""
],
[
"Guler",
"Beyza Nur",
""
],
[
"Bhaduri",
"Indrani",
""
],
[
"Kimbangu",
"Rodney",
""
],
[
"Llerena",
"Nils Ever Murrugarra",
""
],
[
"Shin",
"Hayoung",
""
],
[
"Virguez",
"Lilianny",
""
],
[
"Yanque",
"Rosa Paccotacya",
""
],
[
"Mekhaël",
"Thomas",
""
],
[
"Munoriyarwa",
"Allen",
""
],
[
"Salgado",
"Leslie",
""
],
[
"Basu",
"Debarati",
""
],
[
"Mapaling",
"Curwyn",
""
],
[
"Perez",
"Natalie",
""
],
[
"Gaudet",
"Yves",
""
],
[
"Larrondo",
"Paula",
""
]
] |
2506.09186 | Aaron Hurst | Aaron Hurst, Andrey V. Kalinichev, Klaus Koren and Daniel E. Lucani | Not all those who drift are lost: Drift correction and calibration scheduling for the IoT | null | null | null | null | eess.SP cs.DB | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Sensors provide a vital source of data that link digital systems with the physical world. However, as sensors age, the relationship between what they measure and what they output changes. This is known as sensor drift and poses a significant challenge that, combined with limited opportunity for re-calibration, can severely limit data quality over time. Previous approaches to drift correction typically require large volumes of ground truth data and do not consider measurement or prediction uncertainty. In this paper, we propose a probabilistic sensor drift correction method that takes a fundamental approach to modelling the sensor response using Gaussian Process Regression. Tested using dissolved oxygen sensors, our method delivers mean squared error (MSE) reductions of up to 90% and more than 20% on average. We also propose a novel uncertainty-driven calibration schedule optimisation approach that builds on top of drift correction and further reduces MSE by up to 15.7%. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:04:19 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Hurst",
"Aaron",
""
],
[
"Kalinichev",
"Andrey V.",
""
],
[
"Koren",
"Klaus",
""
],
[
"Lucani",
"Daniel E.",
""
]
] |
2506.09187 | Ahmed Aboudonia | Ahmed Aboudonia, Johannes Estermann, Keith Moffat, Manfred Morari, John Lygeros | A Data-driven Predictive Control Architecture for Train Thermal Energy Management | null | null | null | null | eess.SY cs.SY | http://creativecommons.org/licenses/by/4.0/ | We aim to improve the energy efficiency of train climate control architectures, with a focus on a specific class of regional trains operating throughout Switzerland, especially in Zurich and Geneva. Heating, Ventilation, and Air Conditioning (HVAC) systems represent the second largest energy consumer in these trains after traction. The current architecture comprises a high-level rule-based controller and a low-level tracking controller. To improve train energy efficiency, we propose adding a middle data-driven predictive control layer aimed at minimizing HVAC energy consumption while maintaining passenger comfort. The scheme incorporates a multistep prediction model developed using real-world data collected from a limited number of train coaches. To validate the effectiveness of the proposed architecture, we conduct multiple experiments on a separate set of train coaches; our results suggest energy savings between 10% and 35% with respect to the current architecture. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:07:30 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Aboudonia",
"Ahmed",
""
],
[
"Estermann",
"Johannes",
""
],
[
"Moffat",
"Keith",
""
],
[
"Morari",
"Manfred",
""
],
[
"Lygeros",
"John",
""
]
] |
2506.09188 | Alec McClean | Alec McClean, Alexander W. Levis, Nicholas Williams, and Ivan Diaz | Longitudinal weighted and trimmed treatment effects with flip interventions | null | null | null | null | stat.ME | http://creativecommons.org/licenses/by/4.0/ | Weighting and trimming are popular methods for addressing positivity violations in causal inference. While well-studied with single-timepoint data, standard methods do not easily generalize to address non-baseline positivity violations in longitudinal data, and remain vulnerable to such violations. In this paper, we extend weighting and trimming to longitudinal data via stochastic ``flip'' interventions, which maintain the treatment status of subjects who would have received the target treatment, and flip others' treatment to the target with probability equal to their weight (e.g., overlap weight, trimming indicator). We first show, in single-timepoint data, that flip interventions yield a large class of weighted average treatment effects, ascribing a novel policy interpretation to these popular weighted estimands. With longitudinal data, we then show that flip interventions provide interpretable weighting or trimming on non-baseline covariates and, crucially, yield effects that are identifiable under arbitrary positivity violations. Moreover, we demonstrate that flip interventions are policy-relevant since they could be implemented in practice. By contrast, we show that alternative approaches for weighting on non-baseline covariates fail to achieve this property. We derive flexible and efficient estimators based on efficient influence functions when the weight is a smooth function of the propensity score. Namely, we construct multiply robust-style and sequentially doubly robust-style estimators that achieve root-n consistency and asymptotic normality under nonparametric conditions. Finally, we demonstrate our methods through an analysis of the effect of union membership on earnings. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:12:21 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"McClean",
"Alec",
""
],
[
"Levis",
"Alexander W.",
""
],
[
"Williams",
"Nicholas",
""
],
[
"Diaz",
"Ivan",
""
]
] |
2506.09189 | Esteban Guti\'errez | Esteban Guti\'errez, Rodrigo C\'adiz, Carlos Sing Long, Frederic Font, Xavier Serra | Fractional Fourier Sound Synthesis | Accepted to the International Computer Music Conference (ICMC) 2025 held in Boston, USA. 6 pages and 2 figures | null | null | null | cs.SD eess.AS | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This paper explores the innovative application of the Fractional Fourier Transform (FrFT) in sound synthesis, highlighting its potential to redefine time-frequency analysis in audio processing. As an extension of the classical Fourier Transform, the FrFT introduces fractional order parameters, enabling a continuous interpolation between time and frequency domains and unlocking unprecedented flexibility in signal manipulation. Crucially, the FrFT also opens the possibility of directly synthesizing sounds in the alpha-domain, providing a unique framework for creating timbral and dynamic characteristics unattainable through conventional methods. This work delves into the mathematical principles of the FrFT, its historical evolution, and its capabilities for synthesizing complex audio textures. Through experimental analyses, we showcase novel sound design techniques, such as alpha-synthesis and alpha-filtering, which leverage the FrFT's time-frequency rotation properties to produce innovative sonic results. The findings affirm the FrFT's value as a transformative tool for composers, sound designers, and researchers seeking to push the boundaries of auditory creativity. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:12:41 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Gutiérrez",
"Esteban",
""
],
[
"Cádiz",
"Rodrigo",
""
],
[
"Long",
"Carlos Sing",
""
],
[
"Font",
"Frederic",
""
],
[
"Serra",
"Xavier",
""
]
] |
2506.09190 | Vahid Taghiloo | V.Taghiloo, M.H.Vahidinia | Generalized Symmetries in Shallow Water | 9 pages, 2 columns | null | null | null | hep-th cond-mat.other math-ph math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent developments have extended the concept of global symmetries in several directions, offering new perspectives across a wide range of physical systems. This work shows that generalized global symmetries naturally emerge in shallow water systems. In particular, we demonstrate that both global and gauge subsystem symmetries-previously studied primarily in exotic field theories-arise intrinsically in the dynamics of shallow water flows. A central result is that the local conservation of potential vorticity follows directly from the underlying gauge subsystem symmetries, revealing that the classic Kelvin circulation theorem is rooted in these symmetries. Notably, the associated charge algebra forms a Kac-Moody current algebra, with the level determined by the spatial variation of the Coriolis parameter. Beyond gauge symmetries, we also identify global subsystem symmetries, construct the corresponding Noether charges, and explore their potential applications. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:14:12 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Taghiloo",
"V.",
""
],
[
"Vahidinia",
"M. H.",
""
]
] |
2506.09191 | Paolo Malgaretti Mr | A. Baer, P. Malgaretti, K. Hoellring, J. Harting, Ana-Suncana Smith | Zigzagging Diffusion and Non-Standard Transport in Particle-laden Nanopores Under Extreme Confinement | null | null | null | null | cond-mat.soft | http://creativecommons.org/licenses/by/4.0/ | Understanding transport subject to molecular-scale confinement is key to advancing nanofluidics, yet classical hydrodynamic laws often fail at these scales. Here, we study a model system: transport of toluene as a solvent and small fullerenes as model particles confined within alumina slit nanopores using molecular dynamics simulations. We find that toluene organizes into discrete layers whose commensurability with the pore width leads to a striking, non-monotonic, zig-zag dependence of transport coefficients on confinement. This layering drives oscillations not only in solvent diffusivity but also in flow velocity and permeability under pressure-driven conditions, breaking the expected scaling relations between diffusion, viscosity, and flow. Surprisingly, introducing a nanoparticle does not wash out these effects - although the fullerene perturbs local layering, the nanoparticle diffusivity retains a zig-zag dependence on pore width. Our results demonstrate how structural commensurability and interfacial effects dominate transport in nanoconfined liquids, and lead to important deviations from continuum expectations. These findings establish a microscopic basis for size-dependent transport in nanopores and highlight the need for beyond-hydrodynamic models in confined soft matter systems. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:14:43 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Baer",
"A.",
""
],
[
"Malgaretti",
"P.",
""
],
[
"Hoellring",
"K.",
""
],
[
"Harting",
"J.",
""
],
[
"Smith",
"Ana-Suncana",
""
]
] |
2506.09192 | Fraser Gillan | A. Fraser Gillan, Alan Fitzsimmons, Larry Denneau, Robert J. Siverd, Ken W. Smith, John L. Tonry, David R. Young | Dust production rates in Jupiter-family comets II: Trends and population insights from ATLAS photometry of 116 JFCs | Accepted for publication in The Planetary Science Journal (PSJ) | null | null | null | astro-ph.EP | http://creativecommons.org/licenses/by/4.0/ | Jupiter-family comets (JFCs) have orbital periods of less than 20 years and therefore undergo more frequent sublimation compared to other comet populations. The JFCs therefore represent the ideal dynamical population for investigating the dust production rates at high-cadence. We analyzed observations by the Asteroid Terrestrial-impact Last Alert System (ATLAS) of 74 JFCs that reached perihelion in 2022 and 2023. The work contained in this study builds upon our previous work (Gillan et al. 2024), for a total of 116 JFCs over a four-year period. Using the Afrho parameter, we measured the dust production rates of each JFC as a function of heliocentric distance. We found that there remained a clear preference for JFCs to reach their maximum A(0)frho post-perihelion, with 170P/Christensen, 254P/McNaught and P/2020 WJ5 (Lemmon) reaching a maximum A(0)frho between 200-400 days after perihelion. However, all JFCs reached their maximum dust production within 10% of their orbital period relative to perihelion. Fitting A(0)fp as a function of Rh^n, we measured statistically significant differences in the distribution of pre-perihelion and post-perihelion activity index n, with average activity indices of -5.2 +/- 4.5 and -3.2 +/- 2.7 respectively. We derived upper limits for the nuclear radii of comets 444P/WISE-PANSTARRS and 459P/Catalina as Rn \leq 1.5 +/- 0.2 km and Rn \leq 1.7 +/- 0.1 km respectively. We measured six outbursts in comets 97P/Metcalf-Brewington, 99P/Kowal 1, 118P/Shoemaker-Levy 4, 285P/LINEAR and 382P/Larson. From our four years of observing JFC outbursts in the ATLAS data, the average increase in magnitude was - 1.3 +/- 0.8. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:14:49 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Gillan",
"A. Fraser",
""
],
[
"Fitzsimmons",
"Alan",
""
],
[
"Denneau",
"Larry",
""
],
[
"Siverd",
"Robert J.",
""
],
[
"Smith",
"Ken W.",
""
],
[
"Tonry",
"John L.",
""
],
[
"Young",
"David R.",
""
]
] |
2506.09193 | Yilin Zhuang | Yilin Zhuang, Karthik Duraisamy | LaDCast: A Latent Diffusion Model for Medium-Range Ensemble Weather Forecasting | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Accurate probabilistic weather forecasting demands both high accuracy and efficient uncertainty quantification, challenges that overburden both ensemble numerical weather prediction (NWP) and recent machine-learning methods. We introduce LaDCast, the first global latent-diffusion framework for medium-range ensemble forecasting, which generates hourly ensemble forecasts entirely in a learned latent space. An autoencoder compresses high-dimensional ERA5 reanalysis fields into a compact representation, and a transformer-based diffusion model produces sequential latent updates with arbitrary hour initialization. The model incorporates Geometric Rotary Position Embedding (GeoRoPE) to account for the Earth's spherical geometry, a dual-stream attention mechanism for efficient conditioning, and sinusoidal temporal embeddings to capture seasonal patterns. LaDCast achieves deterministic and probabilistic skill close to that of the European Centre for Medium-Range Forecast IFS-ENS, without any explicit perturbations. Notably, LaDCast demonstrates superior performance in tracking rare extreme events such as cyclones, capturing their trajectories more accurately than established models. By operating in latent space, LaDCast reduces storage and compute by orders of magnitude, demonstrating a practical path toward forecasting at kilometer-scale resolution in real time. We open-source our code and models and provide the training and evaluation pipelines at: https://github.com/tonyzyl/ladcast. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:17:14 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Zhuang",
"Yilin",
""
],
[
"Duraisamy",
"Karthik",
""
]
] |
2506.09194 | Emirhan Bilgi\c{c} | Emirhan Bilgi\c{c}, Neslihan Serap \c{S}eng\"or, Nam{\i}k Berk Yalab{\i}k, Yavuz Selim \.I\c{s}ler, Aykut G\"orkem Gelen, Rahmi Elibol | Integration of Contrastive Predictive Coding and Spiking Neural Networks | 4 pages, 5 figures, 1 table. Accepted at the 2025 33rd Signal Processing and Communications Applications Conference (SIU) | null | null | null | eess.SP cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This study examines the integration of Contrastive Predictive Coding (CPC) with Spiking Neural Networks (SNN). While CPC learns the predictive structure of data to generate meaningful representations, SNN mimics the computational processes of biological neural systems over time. In this study, the goal is to develop a predictive coding model with greater biological plausibility by processing inputs and outputs in a spike-based system. The proposed model was tested on the MNIST dataset and achieved a high classification rate in distinguishing positive sequential samples from non-sequential negative samples. The study demonstrates that CPC can be effectively combined with SNN, showing that an SNN trained for classification tasks can also function as an encoding mechanism. Project codes and detailed results can be accessed on our GitHub page: https://github.com/vnd-ogrenme/ongorusel-kodlama/tree/main/CPC_SNN | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:23:08 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Bilgiç",
"Emirhan",
""
],
[
"Şengör",
"Neslihan Serap",
""
],
[
"Yalabık",
"Namık Berk",
""
],
[
"İşler",
"Yavuz Selim",
""
],
[
"Gelen",
"Aykut Görkem",
""
],
[
"Elibol",
"Rahmi",
""
]
] |
2506.09195 | Haoran Peng | Haoran Peng and Ying-Jun Angela Zhang | Graph Attention-based Decentralized Actor-Critic for Dual-Objective Control of Multi-UAV Swarms | null | null | null | null | eess.SP cs.AI cs.MA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This research focuses on optimizing multi-UAV systems with dual objectives: maximizing service coverage as the primary goal while extending battery lifetime as the secondary objective. We propose a Graph Attention-based Decentralized Actor-Critic (GADC) to optimize the dual objectives. The proposed approach leverages a graph attention network to process UAVs' limited local observation and reduce the dimension of the environment states. Subsequently, an actor-double-critic network is developed to manage dual policies for joint objective optimization. The proposed GADC uses a Kullback-Leibler (KL) divergence factor to balance the tradeoff between coverage performance and battery lifetime in the multi-UAV system. We assess the scalability and efficiency of GADC through comprehensive benchmarking against state-of-the-art methods, considering both theory and experimental aspects. Extensive testing in both ideal settings and NVIDIA Sionna's realistic ray tracing environment demonstrates GADC's superior performance. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:30:53 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Peng",
"Haoran",
""
],
[
"Zhang",
"Ying-Jun Angela",
""
]
] |
2506.09196 | Anna Nickolaevna Morozovska | Eugene A. Eliseev, Anna N. Morozovska, Sergei V. Kalinin, Long-Qing Chen, and Venkatraman Gopalan | Tip-Based Proximity Ferroelectric Switching and Piezoelectric Response in Wurtzite Multilayers | 30 pages, including 7 figures and Appendixes | null | null | null | cond-mat.mtrl-sci physics.app-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Proximity ferroelectricity is a novel paradigm for inducing ferroelectricity, where a non-ferroelectric polar material, which is unswitchable with an external field below the dielectric breakdown field, becomes a practically switchable ferroelectric in direct contact with a thin switchable ferroelectric layer. Here, we develop a Landau-Ginzburg-Devonshire approach to study the proximity effect of local piezoelectric response and polarization reversal in wurtzite ferroelectric multilayers under a sharp electrically biased tip. Using finite element modeling we analyze the probe-induced nucleation of nanodomains, the features of local polarization hysteresis loops and coercive fields in the Al1-xScxN/AlN bilayers and three-layers. Similar to the wurtzite multilayers sandwiched between two parallel electrodes, the regimes of "proximity switching" (where the multilayers collectively switch) and the regime of "proximity suppression" (where they collectively do not switch) are the only two possible regimes in the probe-electrode geometry. However, the parameters and asymmetry of the local piezo-response and polarization hysteresis loops depend significantly on the sequence of the layers with respect to the probe. The physical mechanism of the proximity ferroelectricity in the local probe geometry is a depolarizing electric field determined by the polarization of the layers and their relative thickness. The field, whose direction is opposite to the polarization vector in the layer(s) with the larger spontaneous polarization (such as AlN), renormalizes the double-well ferroelectric potential to lower the steepness of the switching barrier in the "otherwise unswitchable" polar layers. Tip-based control of domains in otherwise non-ferroelectric layers using proximity ferroelectricity can provide nanoscale control of domain reversal in memory, actuation, sensing and optical applications. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:34:08 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Eliseev",
"Eugene A.",
""
],
[
"Morozovska",
"Anna N.",
""
],
[
"Kalinin",
"Sergei V.",
""
],
[
"Chen",
"Long-Qing",
""
],
[
"Gopalan",
"Venkatraman",
""
]
] |
2506.09197 | Sushi George | Sushi Anna George, Vinay Joseph | Adaptive Bandwidth Sharing for Optimizing QoE of Real-Time Video | arXiv admin note: text overlap with arXiv:2401.10681 | null | null | null | cs.NI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The concept of spectrum or bandwidth sharing has gained significant global attention as a means to enhance the efficiency of real-time traffic management in wireless networks. Effective bandwidth sharing enables optimal utilization of available resources, reducing congestion and improving QoE for delay-sensitive applications such as real-time video transmission. In this paper, we propose a novel iterative semi-static bandwidth sharing policy that balances the advantages of both static and dynamic sharing approaches. Our approach minimizes the frequency of coordination between network operators while ensuring efficient resource allocation and meeting the stringent QoE demands of real-time traffic. The proposed policy iteratively optimizes both the spectrum sharing between operators and the resource allocation for individual clients. We establish strong theoretical guarantees for the optimality of the proposed policy and prove that it converges to the optimal static sharing policy irrespective of initial conditions or fluctuations in traffic arrival rates. Additionally, we conduct extensive simulations to evaluate the impact of key system parameters - including step size, hyperperiod length, and arrival process dynamics - on the performance of our policy. Our results demonstrate the effectiveness of the proposed approach in achieving near-optimal bandwidth allocation with reduced overhead, making it a practical solution for real-time wireless applications. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:35:00 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"George",
"Sushi Anna",
""
],
[
"Joseph",
"Vinay",
""
]
] |
2506.09198 | Ali Rezaei | Ali Rezaei, Luc Jaulmes, Maria Bahna, Oliver Thomson Brown, Antonio Barbalace | Low-Level and NUMA-Aware Optimization for High-Performance Quantum Simulation | 12 pages, 9 figures, 2 tables, 2 pseudocodes | null | null | null | quant-ph cs.AR | http://creativecommons.org/licenses/by/4.0/ | Scalable classical simulation of quantum circuits is crucial for advancing both quantum algorithm development and hardware validation. In this work, we focus on performance enhancements through meticulous low-level tuning on a single-node system, thereby not only advancing the performance of classical quantum simulations but also laying the groundwork for scalable, heterogeneous implementations that may eventually bridge the gap toward noiseless quantum computing. Although similar efforts in low-level tuning have been reported in the literature, such implementations have not been released as open-source software, thereby impeding independent evaluation and further development. We introduce an open-source, high-performance extension to the QuEST simulator that brings state-of-the-art low-level and NUMA optimizations to modern computers. Our approach emphasizes locality-aware computation and incorporates hardware-specific optimizations such as NUMA-aware memory allocation, thread pinning, AVX-512 vectorization, aggressive loop unrolling, and explicit memory prefetching. Experiments demonstrate significant speedups - 5.5-6.5x for single-qubit gate operations, 4.5x for two-qubit gates, 4x for Random Quantum Circuits (RQC), and 1.8x for Quantum Fourier Transform (QFT), demonstrating that rigorous performance tuning can substantially extend the practical simulation capacity of classical quantum simulators on current hardware. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:35:11 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Rezaei",
"Ali",
""
],
[
"Jaulmes",
"Luc",
""
],
[
"Bahna",
"Maria",
""
],
[
"Brown",
"Oliver Thomson",
""
],
[
"Barbalace",
"Antonio",
""
]
] |
2506.09199 | Jyotikrishna Dass | Hariharan Ramesh and Jyotikrishna Dass | FLoRIST: Singular Value Thresholding for Efficient and Accurate Federated Fine-Tuning of Large Language Models | 21 pages, 12 figures | null | null | null | cs.LG cs.AI cs.DC | http://creativecommons.org/licenses/by/4.0/ | Integrating Low-Rank Adaptation (LoRA) into federated learning offers a promising solution for parameter-efficient fine-tuning of Large Language Models (LLMs) without sharing local data. However, several methods designed for federated LoRA present significant challenges in balancing communication efficiency, model accuracy, and computational cost, particularly among heterogeneous clients. These methods either rely on simplistic averaging of local adapters, which introduces aggregation noise, require transmitting large stacked local adapters, leading to poor communication efficiency, or necessitate reconstructing memory-dense global weight-update matrix and performing computationally expensive decomposition to design client-specific low-rank adapters. In this work, we propose FLoRIST, a federated fine-tuning framework that achieves mathematically accurate aggregation without incurring high communication or computational overhead. Instead of constructing the full global weight-update matrix at the server, FLoRIST employs an efficient decomposition pipeline by performing singular value decomposition on stacked local adapters separately. This approach operates within a compact intermediate space to represent the accumulated information from local LoRAs. We introduce tunable singular value thresholding for server-side optimal rank selection to construct a pair of global low-rank adapters shared by all clients. Extensive empirical evaluations across multiple datasets and LLMs demonstrate that FLoRIST consistently strikes the best balance between superior communication efficiency and competitive performance in both homogeneous and heterogeneous setups. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:36:36 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Ramesh",
"Hariharan",
""
],
[
"Dass",
"Jyotikrishna",
""
]
] |
2506.09201 | Daniella Bardalez Gagliuffi PhD | Daniella C. Bardalez Gagliuffi, William O. Balmer, Laurent Pueyo, Timothy D. Brandt, Mark R. Giovinazzi, Sarah Millholland, Brennen Black, Tiger Lu, Malena Rice, James Mang, Caroline Morley, Brianna Lacy, Julien Girard, Elisabeth Matthews, Aarynn Carter, Brendan P. Bowler, Jacqueline K. Faherty, Clemence Fontanive, and Emily Rickman | JWST Coronagraphic Images of 14 Her c: a Cold Giant Planet in a Dynamically Hot, Multi-planet System | 25 pages, 10 figures | null | null | null | astro-ph.EP astro-ph.SR | http://creativecommons.org/licenses/by/4.0/ | Most observed multi-planet systems are coplanar, in a dynamically "cold" configuration of concentric orbits like our own Solar System. With the James Webb Space Telescope (JWST) we have detected 14 Her c, the first mature and cold exoplanet directly imaged in a dynamically "hot", multi-planet system. With large eccentricities and a nonzero mutual inclination, the present-day architecture of this system points to a turbulent past and ongoing angular momentum exchange between the planetary orbits of 14 Her b and c. The temperature of 14 Her c rivals both the coldest imaged exoplanet and the coldest known brown dwarf. Moreover, its photometry at 4.4 mu is consistent with the presence of carbon disequilibrium chemistry and water ice clouds in its atmosphere. 14 Her c presents a unique laboratory to study giant planet formation, dynamical evolution of multi-planet system architectures, and atmospheric composition and dynamics in extremely cold worlds. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:42:42 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Gagliuffi",
"Daniella C. Bardalez",
""
],
[
"Balmer",
"William O.",
""
],
[
"Pueyo",
"Laurent",
""
],
[
"Brandt",
"Timothy D.",
""
],
[
"Giovinazzi",
"Mark R.",
""
],
[
"Millholland",
"Sarah",
""
],
[
"Black",
"Brennen",
""
],
[
"Lu",
"Tiger",
""
],
[
"Rice",
"Malena",
""
],
[
"Mang",
"James",
""
],
[
"Morley",
"Caroline",
""
],
[
"Lacy",
"Brianna",
""
],
[
"Girard",
"Julien",
""
],
[
"Matthews",
"Elisabeth",
""
],
[
"Carter",
"Aarynn",
""
],
[
"Bowler",
"Brendan P.",
""
],
[
"Faherty",
"Jacqueline K.",
""
],
[
"Fontanive",
"Clemence",
""
],
[
"Rickman",
"Emily",
""
]
] |
2506.09203 | Sumanta Kundu | Sumanta Kundu, Anca Opris, Yosuke Aoki, and Takahiro Hatano | Inter-event time statistics of earthquakes as a gauge of volcano activity | null | null | null | null | physics.geo-ph | http://creativecommons.org/licenses/by/4.0/ | The probability distribution of inter-event time (IET) between two consecutive earthquakes is a measure for the uncertainty in the occurrence time of earthquakes in a region of interest. It is well known that the IET distribution for regular earthquakes is commonly characterized by a power law with the exponent of 0.3. However, less is known about other classes of earthquakes, such as volcanic earthquakes, which do not manifest mainshock-aftershocks sequences. Since volcanic earthquakes are caused by the movement of magmas, their IET distribution may be closely related to the volcanic activities and therefore of particular interest. Nevertheless, the general form of IET distribution for volcanic earthquakes and its dependence on volcanic activity are still unknown. Here we show that the power-law exponent characterizing the IET distribution exhibits a few common values depending on the stage of volcanic activity. Volcanoes with steady seismicity exhibit the lowest exponent ranging from 0.6 to 0.7. During the burst period, when the earthquake rate is highest, the exponent reaches its peak at approximately 1.3. In the preburst phase, the exponent takes on the intermediate value of 1.0. These values are common to several different volcanoes. Since the preburst phase is characterized by the distinct exponent value, it may serve as an indicator of imminent volcanic activity that is accompanied by a surge in seismic events. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:48:58 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Kundu",
"Sumanta",
""
],
[
"Opris",
"Anca",
""
],
[
"Aoki",
"Yosuke",
""
],
[
"Hatano",
"Takahiro",
""
]
] |
2506.09204 | Seyed Sahand Mohammadi Ziabari | Xiaotian Chen, Hongyun Liu, Seyed Sahand Mohammadi Ziabari | A Topological Improvement of the Overall Performance of Sparse Evolutionary Training: Motif-Based Structural Optimization of Sparse MLPs Project | null | null | null | null | cs.NE cs.AI | http://creativecommons.org/licenses/by/4.0/ | Deep Neural Networks (DNNs) have been proven to be exceptionally effective and have been applied across diverse domains within deep learning. However, as DNN models increase in complexity, the demand for reduced computational costs and memory overheads has become increasingly urgent. Sparsity has emerged as a leading approach in this area. The robustness of sparse Multi-layer Perceptrons (MLPs) for supervised feature selection, along with the application of Sparse Evolutionary Training (SET), illustrates the feasibility of reducing computational costs without compromising accuracy. Moreover, it is believed that the SET algorithm can still be improved through a structural optimization method called motif-based optimization, with potential efficiency gains exceeding 40% and a performance decline of under 4%. This research investigates whether the structural optimization of Sparse Evolutionary Training applied to Multi-layer Perceptrons (SET-MLP) can enhance performance and to what extent this improvement can be achieved. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:49:07 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Chen",
"Xiaotian",
""
],
[
"Liu",
"Hongyun",
""
],
[
"Ziabari",
"Seyed Sahand Mohammadi",
""
]
] |
2506.09205 | Haiyan Wang | Haiyan Wang | Genetic Transformer-Assisted Quantum Neural Networks for Optimal Circuit Design | null | null | null | null | quant-ph cs.NE | http://creativecommons.org/licenses/by/4.0/ | We introduce Genetic Transformer Assisted Quantum Neural Networks (GTQNNs), a hybrid learning framework that combines a transformer encoder with a shallow variational quantum circuit and automatically fine tunes the circuit via the NSGA-II multi objective genetic algorithm. The transformer reduces high-dimensional classical data to a compact, qubit sized representation, while NSGA-II searches for Pareto optimal circuits that (i) maximize classification accuracy and (ii) minimize primitive gate count an essential constraint for noisy intermediate-scale quantum (NISQ) hardware. Experiments on four benchmarks (Iris, Breast Cancer, MNIST, and Heart Disease) show that GTQNNs match or exceed state of the art quantum models while requiring much fewer gates for most cases. A hybrid Fisher information analysis further reveals that the trained networks operate far from barren plateaus; the leading curvature directions increasingly align with the quantum subspace as the qubit budget grows, confirming that the transformer front end has effectively condensed the data. Together, these results demonstrate that GTQNNs deliver competitive performance with a quantum resource budget well suited to present-day NISQ devices. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:49:21 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Wang",
"Haiyan",
""
]
] |
2506.09206 | Ahmed Attia | Ahmed Adel Attia, Jing Liu, Carl Espy-Wilson | SimClass: A Classroom Speech Dataset Generated via Game Engine Simulation For Automatic Speech Recognition Research | null | null | null | null | cs.SD cs.AI cs.CL eess.AS | http://creativecommons.org/licenses/by/4.0/ | The scarcity of large-scale classroom speech data has hindered the development of AI-driven speech models for education. Public classroom datasets remain limited, and the lack of a dedicated classroom noise corpus prevents the use of standard data augmentation techniques.
In this paper, we introduce a scalable methodology for synthesizing classroom noise using game engines, a framework that extends to other domains. Using this methodology, we present SimClass, a dataset that includes both a synthesized classroom noise corpus and a simulated classroom speech dataset. The speech data is generated by pairing a public children's speech corpus with YouTube lecture videos to approximate real classroom interactions in clean conditions. Our experiments on clean and noisy speech demonstrate that SimClass closely approximates real classroom speech, making it a valuable resource for developing robust speech recognition and enhancement models. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:51:57 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Attia",
"Ahmed Adel",
""
],
[
"Liu",
"Jing",
""
],
[
"Espy-Wilson",
"Carl",
""
]
] |
2506.09208 | Anru R. Zhang | Jianbin Tan and Yan Zhang and Chuan Hong and T. Tony Cai and Tianxi Cai and Anru R. Zhang | Integrated Analysis for Electronic Health Records with Structured and Sporadic Missingness | null | null | null | null | stat.AP stat.CO stat.ME stat.ML | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Objectives: We propose a novel imputation method tailored for Electronic Health Records (EHRs) with structured and sporadic missingness. Such missingness frequently arises in the integration of heterogeneous EHR datasets for downstream clinical applications. By addressing these gaps, our method provides a practical solution for integrated analysis, enhancing data utility and advancing the understanding of population health.
Materials and Methods: We begin by demonstrating structured and sporadic missing mechanisms in the integrated analysis of EHR data. Following this, we introduce a novel imputation framework, Macomss, specifically designed to handle structurally and heterogeneously occurring missing data. We establish theoretical guarantees for Macomss, ensuring its robustness in preserving the integrity and reliability of integrated analyses. To assess its empirical performance, we conduct extensive simulation studies that replicate the complex missingness patterns observed in real-world EHR systems, complemented by validation using EHR datasets from the Duke University Health System (DUHS).
Results: Simulation studies show that our approach consistently outperforms existing imputation methods. Using datasets from three hospitals within DUHS, Macomss achieves the lowest imputation errors for missing data in most cases and provides superior or comparable downstream prediction performance compared to benchmark methods.
Conclusions: We provide a theoretically guaranteed and practically meaningful method for imputing structured and sporadic missing data, enabling accurate and reliable integrated analysis across multiple EHR datasets. The proposed approach holds significant potential for advancing research in population health. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:59:49 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Tan",
"Jianbin",
""
],
[
"Zhang",
"Yan",
""
],
[
"Hong",
"Chuan",
""
],
[
"Cai",
"T. Tony",
""
],
[
"Cai",
"Tianxi",
""
],
[
"Zhang",
"Anru R.",
""
]
] |
2506.09209 | Leandro Anghinoni | Leandro Anghinoni, Pablo Zivic, Jorge Adrian Sanchez | Revisiting Graph Projections for Effective Complementary Product Recommendation | null | null | null | null | cs.IR cs.LG | http://creativecommons.org/licenses/by-sa/4.0/ | Complementary product recommendation is a powerful strategy to improve customer experience and retail sales. However, recommending the right product is not a simple task because of the noisy and sparse nature of user-item interactions. In this work, we propose a simple yet effective method to predict a list of complementary products given a query item, based on the structure of a directed weighted graph projected from the user-item bipartite graph. We revisit bipartite graph projections for recommender systems and propose a novel approach for inferring complementarity relationships from historical user-item interactions. We compare our model with recent methods from the literature and show, despite the simplicity of our approach, an average improvement of +43% and +38% over sequential and graph-based recommenders, respectively, over different benchmarks. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:59:49 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Anghinoni",
"Leandro",
""
],
[
"Zivic",
"Pablo",
""
],
[
"Sanchez",
"Jorge Adrian",
""
]
] |
2506.09210 | Alejandro Ram\'irez-Ya\~nez | Alejandro Ram\'irez-Ya\~nez, Thomas Gorin, Rodrigo A. Vicencio and V\'ictor A. Gopar | Anomalous localization of light in one-dimensional L\'evy photonic lattices | Includes supplemental material | null | null | null | cond-mat.dis-nn physics.optics | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Localization of coherent propagating waves has been extensively studied over the years, primarily in homogeneous random media. However, significantly less attention has been given to wave localization in inhomogeneous systems, where the standard picture of Anderson localization does not apply, as we demonstrate here. We fabricate photonic lattices with inhomogeneous disorder, modeled by heavy-tailed $\alpha$-stable distributions, and measure the output light intensity profiles. We demonstrate that the spatial localization of light is described by a stretched exponential function, with a stretching parameter $\alpha$, and an asymmetric localized profile with respect to the excitation site. We support our experimental and theoretical findings with extensive tight-binding simulations. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 19:59:55 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Ramírez-Yañez",
"Alejandro",
""
],
[
"Gorin",
"Thomas",
""
],
[
"Vicencio",
"Rodrigo A.",
""
],
[
"Gopar",
"Víctor A.",
""
]
] |
2506.09211 | Jemima Tabeart | I. Dau\v{z}ickait\.e, M. A. Freitag, S. G\"urol, A. S. Lawless, A. Ramage, J. A. Scott, J. M. Tabeart | An Introduction to Solving the Least-Squares Problem in Variational Data Assimilation | null | null | null | null | math.NA cs.NA | http://creativecommons.org/licenses/by/4.0/ | Variational data assimilation is a technique for combining measured data with dynamical models. It is a key component of Earth system state estimation and is commonly used in weather and ocean forecasting. The approach involves a large-scale generalized nonlinear least-squares problem. Solving the resulting sequence of sparse linear subproblems requires the use of sophisticated numerical linear algebra methods. In practical applications, the computational demands severely limit the number of iterations of a Krylov subspace solver that can be performed and so high-quality preconditioners are vital. In this paper, we introduce variational data assimilation from a numerical linear algebra perspective and review current solution techniques, with a focus on the challenges that arise in large-scale geophysical systems. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:02:51 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Daužickaitė",
"I.",
""
],
[
"Freitag",
"M. A.",
""
],
[
"Gürol",
"S.",
""
],
[
"Lawless",
"A. S.",
""
],
[
"Ramage",
"A.",
""
],
[
"Scott",
"J. A.",
""
],
[
"Tabeart",
"J. M.",
""
]
] |
2506.09212 | Lucas Joos | Lucas Joos, Gavin J. Mooney, Maximilian T. Fischer, Daniel A. Keim, Falk Schreiber, Helen C. Purchase and Karsten Klein | Show Me Your Best Side: Characteristics of User-Preferred Perspectives for 3D Graph Drawings | null | null | null | null | cs.HC | http://creativecommons.org/licenses/by/4.0/ | The visual analysis of graphs in 3D has become increasingly popular, accelerated by the rise of immersive technology, such as augmented and virtual reality. Unlike 2D drawings, 3D graph layouts are highly viewpoint-dependent, making perspective selection critical for revealing structural and relational patterns. Despite its importance, there is limited empirical evidence guiding what constitutes an effective or preferred viewpoint from the user's perspective. In this paper, we present a systematic investigation into user-preferred viewpoints in 3D graph visualisations. We conducted a controlled study with 23 participants in a virtual reality environment, where users selected their most and least preferred viewpoints for 36 different graphs varying in size and layout. From this data, enriched by qualitative feedback, we distil common strategies underlying viewpoint choice. We further analyse the alignment of user preferences with classical 2D aesthetic criteria (e.g., Crossings), 3D-specific measures (e.g., Node-Node Occlusion), and introduce a novel measure capturing the perceivability of a graph's principal axes (Isometric Viewpoint Deviation). Our data-driven analysis indicates that Stress, Crossings, Gabriel Ratio, Edge-Node Overlap, and Isometric Viewpoint Deviation are key indicators of viewpoint preference. Beyond our findings, we contribute a publicly available dataset consisting of the graphs and computed aesthetic measures, supporting further research and the development of viewpoint evaluation measures for 3D graph drawing. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:13:19 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Joos",
"Lucas",
""
],
[
"Mooney",
"Gavin J.",
""
],
[
"Fischer",
"Maximilian T.",
""
],
[
"Keim",
"Daniel A.",
""
],
[
"Schreiber",
"Falk",
""
],
[
"Purchase",
"Helen C.",
""
],
[
"Klein",
"Karsten",
""
]
] |
2506.09213 | Dafni Tsipa | D. Tsipa | On the linearity of fundamental groups of rose and star graphs of groups | 27 pages | null | null | null | math.GR | http://creativecommons.org/licenses/by/4.0/ | Let $G$ be a fundamental group of a graph of group where the graph is a rose or a star graph and the vertex groups are free groups, free abelian groups or right-angled Artin groups. We prove the linearity of $G$ over $\mathbb{Z}$ under various conditions on its edge groups. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:13:52 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Tsipa",
"D.",
""
]
] |
2506.09214 | Takuya Hatomura | Takuya Hatomura | Classical algorithm inspired by the feedback-based algorithm for quantum optimization and local counterdiabatic driving | null | null | null | null | quant-ph cond-mat.stat-mech | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We propose a quantum-inspired classical algorithm for combinatorial optimization problems, named the counterdiabaticity-assisted classical algorithm for optimization (CACAO). In this algorithm, a solution of a given combinatorial optimization problem is heuristically searched with classical spin dynamics based on quantum Lyapunov control of local counterdiabatic driving. We compare the performance of CACAO with that of quantum time-evolution algorithms, i.e., quantum annealing, the feedback-based algorithm for quantum optimization (known as FALQON), and the counterdiabatic feedback-based quantum algorithm (known as CD-FQA). We also study the performance of CACAO applied to large systems up to $10,000$ spins. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:16:44 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Hatomura",
"Takuya",
""
]
] |
2506.09215 | Greyson Brothers | Greyson Brothers | Robust Noise Attenuation via Adaptive Pooling of Transformer Outputs | [ICML 2025 Spotlight Poster] To be published in the Forty-Second International Conference on Machine Learning (ICML) Proceedings | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | We investigate the design of pooling methods used to summarize the outputs of transformer embedding models, primarily motivated by reinforcement learning and vision applications. This work considers problems where a subset of the input vectors contains requisite information for a downstream task (signal) while the rest are distractors (noise). By framing pooling as vector quantization with the goal of minimizing signal loss, we demonstrate that the standard methods used to aggregate transformer outputs, AvgPool, MaxPool, and ClsToken, are vulnerable to performance collapse as the signal-to-noise ratio (SNR) of inputs fluctuates. We then show that an attention-based adaptive pooling method can approximate the signal-optimal vector quantizer within derived error bounds for any SNR. Our theoretical results are first validated by supervised experiments on a synthetic dataset designed to isolate the SNR problem, then generalized to standard relational reasoning, multi-agent reinforcement learning, and vision benchmarks with noisy observations, where transformers with adaptive pooling display superior robustness across tasks. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:18:32 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Brothers",
"Greyson",
""
]
] |
2506.09216 | Qing Xia | Qing (Nancy) Xia, Advait Sarkar, Duncan Brumby, Anna Cox | "How do you even know that stuff?": Barriers to expertise sharing among spreadsheet users | Accepted at CSCW 2025 | null | null | null | cs.HC cs.CY | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Spreadsheet collaboration provides valuable opportunities for learning and expertise sharing between colleagues. Sharing expertise is essential for the retention of important technical skillsets within organisations, but previous studies suggest that spreadsheet experts often fail to disseminate their knowledge to others. We suggest that social norms and beliefs surrounding the value of spreadsheet use significantly influence user engagement in sharing behaviours. To explore this, we conducted 31 semi-structured interviews with professional spreadsheet users from two separate samples. We found that spreadsheet providers face challenges in adapting highly personalised strategies to often subjective standards and evaluating the appropriate social timing of sharing. In addition, conflicted self-evaluations of one's spreadsheet expertise, dismissive normative beliefs about the value of this knowledge, and concerns about the potential disruptions associated with collaboration can further deter sharing. We suggest these observations reflect the challenges of long-term learning in feature-rich software designed primarily with initial learnability in mind. We therefore provide implications for design to navigate this tension. Overall, our findings demonstrate how the complex interaction between technology design and social dynamics can shape collaborative learning behaviours in the context of feature-rich software. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:21:29 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Qing",
"",
"",
"Nancy"
],
[
"Xia",
"",
""
],
[
"Sarkar",
"Advait",
""
],
[
"Brumby",
"Duncan",
""
],
[
"Cox",
"Anna",
""
]
] |
2506.09217 | Liang Shi | Boyu Jiang, Liang Shi, Zhengzhi Lin, Loren Stowe, Feng Guo | Perception Characteristics Distance: Measuring Stability and Robustness of Perception System in Dynamic Conditions under a Certain Decision Rule | null | null | null | null | cs.RO cs.CV stat.AP | http://creativecommons.org/licenses/by/4.0/ | The performance of perception systems in autonomous driving systems (ADS) is strongly influenced by object distance, scene dynamics, and environmental conditions such as weather. AI-based perception outputs are inherently stochastic, with variability driven by these external factors, while traditional evaluation metrics remain static and event-independent, failing to capture fluctuations in confidence over time. In this work, we introduce the Perception Characteristics Distance (PCD) -- a novel evaluation metric that quantifies the farthest distance at which an object can be reliably detected, incorporating uncertainty in model outputs. To support this, we present the SensorRainFall dataset, collected on the Virginia Smart Road using a sensor-equipped vehicle (cameras, radar, LiDAR) under controlled daylight-clear and daylight-rain scenarios, with precise ground-truth distances to the target objects. Statistical analysis reveals the presence of change points in the variance of detection confidence score with distance. By averaging the PCD values across a range of detection quality thresholds and probabilistic thresholds, we compute the mean PCD (mPCD), which captures the overall perception characteristics of a system with respect to detection distance. Applying state-of-the-art perception models shows that mPCD captures meaningful reliability differences under varying weather conditions -- differences that static metrics overlook. PCD provides a principled, distribution-aware measure of perception performance, supporting safer and more robust ADS operation, while the SensorRainFall dataset offers a valuable benchmark for evaluation. The SensorRainFall dataset is publicly available at https://www.kaggle.com/datasets/datadrivenwheels/sensorrainfall, and the evaluation code is open-sourced at https://github.com/datadrivenwheels/PCD_Python. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:22:09 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Jiang",
"Boyu",
""
],
[
"Shi",
"Liang",
""
],
[
"Lin",
"Zhengzhi",
""
],
[
"Stowe",
"Loren",
""
],
[
"Guo",
"Feng",
""
]
] |
2506.09218 | Bruno Ferenc \v{S}egedin | Bruno Ferenc \v{S}egedin | A Technique for Isolating Lexically-Independent Phonetic Dependencies in Generative CNNs | null | null | null | null | cs.CL cs.SD eess.AS | http://creativecommons.org/licenses/by/4.0/ | The ability of deep neural networks (DNNs) to represent phonotactic generalizations derived from lexical learning remains an open question. This study (1) investigates the lexically-invariant generalization capacity of generative convolutional neural networks (CNNs) trained on raw audio waveforms of lexical items and (2) explores the consequences of shrinking the fully-connected layer (FC) bottleneck from 1024 channels to 8 before training. Ultimately, a novel technique for probing a model's lexically-independent generalizations is proposed that works only under the narrow FC bottleneck: generating audio outputs by bypassing the FC and inputting randomized feature maps into the convolutional block. These outputs are equally biased by a phonotactic restriction in training as are outputs generated with the FC. This result shows that the convolutional layers can dynamically generalize phonetic dependencies beyond lexically-constrained configurations learned by the FC. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:22:33 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Šegedin",
"Bruno Ferenc",
""
]
] |
2506.09219 | Rosine Lallement | R. Lallement | Diffuse Interstellar bands and ultraviolet extinction bump: a Milky Way perspective on distant galaxies | 6 pages, 5 figures, to appear in Astronomy Astrophysics | null | null | null | astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | The spectral width wB and the center wavelength lB of the UV absorption bump uvB measured for two z=7 galaxies were found to differ from Milky Way (MW) values. A decrease of wB by 45% and a positive shift of lB by 70-80 Ang were measured. In the MW, the uvB amplitude hB and wB do vary; however, such a narrow bump has never been observed and no variability of lB has been convincingly found. Recently, links have been found between both hB and wB and the strength of several Diffuse Interstellar absorption Bands (DIBs). They were found to be limited to the sigma-type DIBs and their detection to be strongly favored if the data were limited to monocloud-type lines-of-sight, selected based on 3D dust maps. We extended the study of the links between MW DIBs and uvB to the lB value and to the hB to continuum ratio, and compared MW variations of the bump parameters to high z values. We used the 5780 and 6284 A sigma, and 5797 and 5850 A zeta DIBs. Similarly to hB and wB, lB reacts to the abundance of sigma carriers and is insensitive to the abundance of zeta DIB carriers, which dominate in dense cloud cores. A strong abundance of sigma carriers induces a shift of lB to longer wavelengths and a decrease of wB. The MW lB and wB variability range is about half the difference between average MW values and values in the distant galaxies. These results reinforce the hypothesis of the existence of individual types of hydrocarbon molecules responsible for both DIBs and part of the UV bump. They show that the majority of species responsible for narrow and positively shifted bumps in distant galaxies have a link with or are those producing the sigma DIBs and the long-wavelength part of the bump in the MW, and that species producing the short-wavelength part of the bump in the MW are of a different nature and absent along the paths to regions of those distant galaxies that contribute most to the UV emission. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:23:40 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Lallement",
"R.",
""
]
] |
2506.09220 | Tawfiq Ammari | Karen Joy, Tawfiq Ammari, Alyssa Sheehan | Beyond the Hype: Mapping Uncertainty and Gratification in AI Assistant Use | null | null | null | null | cs.HC | http://creativecommons.org/licenses/by/4.0/ | This paper examines the gap between the promises and real-world performance of emerging AI personal assistants. Drawing on interviews with early adopters of devices like Rabbit R1 and Humane AI Pin, as well as services like Ohai and Docus, we map user experiences through the lens of Uses and Gratifications and Uncertainty Reduction Theory. We identify three core types of user uncertainty, functional, interactional, and social, and explore how each disrupts different user gratifications. We show that while marketing hype fuels initial adoption, unmet expectations often result in frustration or abandonment. Our findings highlight the importance of transparency, task-specific design, and user control over contextual memory and personalization. We provide design and policy recommendations, including user-facing explainability tools and calls for regulatory benchmarks such as CI Bench, to guide ethical and interpretable AI integration. Our study offers actionable insights for creating more usable, trustworthy, and socially aligned AI assistants. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:24:32 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Joy",
"Karen",
""
],
[
"Ammari",
"Tawfiq",
""
],
[
"Sheehan",
"Alyssa",
""
]
] |
2506.09221 | Michael Soprano | Michael Soprano | In Crowd Veritas: Leveraging Human Intelligence To Fight Misinformation | PhD thesis, University of Udine, defended May 2023, 458 pages | null | null | null | cs.IR cs.CY cs.SI | http://creativecommons.org/licenses/by/4.0/ | The spread of online misinformation poses serious threats to democratic societies. Traditionally, expert fact-checkers verify the truthfulness of information through investigative processes. However, the volume and immediacy of online content present major scalability challenges. Crowdsourcing offers a promising alternative by leveraging non-expert judgments, but it introduces concerns about bias, accuracy, and interpretability. This thesis investigates how human intelligence can be harnessed to assess the truthfulness of online information, focusing on three areas: misinformation assessment, cognitive biases, and automated fact-checking systems. Through large-scale crowdsourcing experiments and statistical modeling, it identifies key factors influencing human judgments and introduces a model for the joint prediction and explanation of truthfulness. The findings show that non-expert judgments often align with expert assessments, particularly when factors such as timing and experience are considered. By deepening our understanding of human judgment and bias in truthfulness assessment, this thesis contributes to the development of more transparent, trustworthy, and interpretable systems for combating misinformation. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:25:10 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Soprano",
"Michael",
""
]
] |
2506.09222 | Aleksandr Artemev | Aleksandr Artemev | $x-y$ swap for $(2,2p+1)$ minimal string | 31 pages | null | null | null | hep-th math-ph math.MP | http://creativecommons.org/licenses/by/4.0/ | We continue the study of 2D gravity -- ``matrix model'' duality on the example of $(2,2p+1)$ minimal string. We propose a reformulation of the duality, related to a more conventional one by ``$x-y$ swap'' in the language of topological recursion. This formulation elucidates some conceptual and technical difficulties in the dictionary of the duality and relation to other examples. In particular, it allows to circumvent the necessity to use ``resonance transformations'', that were previously introduced to match the worldsheet and ``matrix model'' correlators, and the expressions for minimal string amplitudes in this approach are reminiscent of the ones obtained recently for ``complex Liouville string'' theory. Using this new approach, we also formulate a conjecture on how one can compute amplitudes with operators other than tachyons in the dual theory. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:26:36 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Artemev",
"Aleksandr",
""
]
] |
2506.09223 | Diego Goldsztajn | Konstantin Avrachenkov and Diego Goldsztajn | Semi-supervised Community Detection using Glauber Dynamics for an Ising Model | 40 pages, 3 figures | null | null | null | math.ST stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider graphs with two communities and analyze an algorithm for learning the community labels when the edges of the graph and only a small fraction of the labels are known in advance. The algorithm is based on the Glauber dynamics for an Ising model where the energy function includes a quadratic penalty on the magnetization. The analysis focuses on graphs sampled from a Stochastic Block Model (SBM) with slowly growing mean degree. We derive a mean-field limit for the magnetization of each community, which can be used to choose the run-time of the algorithm to obtain a target accuracy level. We further prove that almost exact recovery is achieved in a number of iterations that is quasi-linear in the number of nodes. As a special case, our results provide the first rigorous analysis of the label propagation algorithm in the SBM with slowly diverging mean degree. We complement our theoretical results with several numerical experiments. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:28:44 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Avrachenkov",
"Konstantin",
""
],
[
"Goldsztajn",
"Diego",
""
]
] |
2506.09224 | David Halpern | David Halpern and Hsien-Hung Wei | Moving contact lines of power-law fluids: How nonlinear fluid rheology drastically alters stress singularity and dynamic wetting behavior | 36 pages, 5 figures | null | null | null | physics.flu-dyn | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Power-law fluids can strongly affect the degree of the contact line stress singularity and hence the nature of moving contact lines. We develop a framework beyond the classical paradigm for power-law fluids, providing a unified account for the distinct behaviors of the advancing contact lines. We show that the apparent dynamic contact angle $\theta_d$ can depend on the extent of the characteristic dissipation length $h^* \propto U_n/(n-1)$, altering its dependence on the contact line speed $U$. For shear-thinning fluids, we find $\theta_d \sim (h/h^*)^{(1-n)/3}$, with contact line motion being dissipated within $h^*$ extending beyond the local wedge height $h$ without requiring a cutoff. In drop spreading problems, $\theta_d$ varies with the spreading radius $R$, leading to $\theta_d \propto U^{3n/(2n+7)}$ consistent with the spreading law $R \propto t^{n/(3n+7)}$ derived from a self-similar solution, where $R$ is the spreading radius and $t$ is time. For shear-thickening fluids, the apparent contact line motion is characterized by $\theta_d \sim (h^*/h_m)^{(1-n)/3}$, where dissipation is concentrated within $h^*$ which is smaller than the microscopic liquid height $h_m$ near the contact line. In fact, the dynamic contact angle relationship in this case can be expressed as the Cox-Voinov law $\theta_d \sim Ca_{eff}^{1/3}$ in terms of a capillary number $Ca_{eff} =\eta_f U/\gamma$ where $\gamma$ is the surface tension and $\eta_f \propto (U/ h_m)^{n-1}$ is the viscosity based on the local shear rate $U/h_m$ across $h_m$. We also show that a precursor film induced by molecular forces ahead of the wedge leads to $h_m \propto U^{-n/(4-n)}$ and hence $\theta_d \propto U^{3n/(4-n)}$, making the spreading behavior highly sensitive to the contact line microstructure. Our predictions show good agreement with experimental results. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:30:21 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Halpern",
"David",
""
],
[
"Wei",
"Hsien-Hung",
""
]
] |
2506.09225 | Hao Jiang | Hao Jiang and Zhaolin Wang and Yue Liu and Hyundong Shin and Arumugam Nallanathan and Yuanwei Liu | Near-Field Sensing Enabled Predictive Beamforming: Fundamentals, Framework, and Opportunities | Submitted to IEEE | null | null | null | eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The article proposes a novel near-field predictive beamforming framework for high-mobility wireless networks. Specifically, due to the spherical waves and non-uniform Doppler frequencies brought by the near-field region, the new ability of full-dimensional location and velocity sensing is characterized. Building on this foundation, the near-field predictive beamforming framework is proposed to proactively design beamformers for mobility users following arbitrary trajectories. Compared to the conventional far-field counterpart, the near-field predictive beamforming stands out due to: i) Prior-Knowledge-Free Prediction, and ii) Low-Complexity and Generalizable System Design. To realize these advantages, the implementation methods are discussed, followed by a case study confirming the benefits of the proposed framework. Finally, the article highlights promising research opportunities inspired by the proposed framework. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:30:27 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Jiang",
"Hao",
""
],
[
"Wang",
"Zhaolin",
""
],
[
"Liu",
"Yue",
""
],
[
"Shin",
"Hyundong",
""
],
[
"Nallanathan",
"Arumugam",
""
],
[
"Liu",
"Yuanwei",
""
]
] |
2506.09226 | Rathijit Sen | Bowen Wu, Wei Cui, Carlo Curino, Matteo Interlandi, Rathijit Sen | Terabyte-Scale Analytics in the Blink of an Eye | null | null | null | null | cs.DB cs.DC cs.PF | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | For the past two decades, the DB community has devoted substantial research to take advantage of cheap clusters of machines for distributed data analytics -- we believe that we are at the beginning of a paradigm shift. The scaling laws and popularity of AI models lead to the deployment of incredibly powerful GPU clusters in commercial data centers. Compared to CPU-only solutions, these clusters deliver impressive improvements in per-node compute, memory bandwidth, and inter-node interconnect performance. In this paper, we study the problem of scaling analytical SQL queries on distributed clusters of GPUs, with the stated goal of establishing an upper bound on the likely performance gains. To do so, we build a prototype designed to maximize performance by leveraging ML/HPC best practices, such as group communication primitives for cross-device data movements. This allows us to conduct thorough performance experimentation to point our community towards a massive performance opportunity of at least 60$\times$. To make these gains more relatable, before you can blink twice, our system can run all 22 queries of TPC-H at a 1TB scale factor! | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:30:31 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Wu",
"Bowen",
""
],
[
"Cui",
"Wei",
""
],
[
"Curino",
"Carlo",
""
],
[
"Interlandi",
"Matteo",
""
],
[
"Sen",
"Rathijit",
""
]
] |
2506.09227 | Jie Ren | Jie Ren, Yue Xing, Yingqian Cui, Charu C. Aggarwal, Hui Liu | SoK: Machine Unlearning for Large Language Models | null | null | null | null | cs.LG cs.CR | http://creativecommons.org/licenses/by/4.0/ | Large language model (LLM) unlearning has become a critical topic in machine learning, aiming to eliminate the influence of specific training data or knowledge without retraining the model from scratch. A variety of techniques have been proposed, including Gradient Ascent, model editing, and re-steering hidden representations. While existing surveys often organize these methods by their technical characteristics, such classifications tend to overlook a more fundamental dimension: the underlying intention of unlearning--whether it seeks to truly remove internal knowledge or merely suppress its behavioral effects. In this SoK paper, we propose a new taxonomy based on this intention-oriented perspective. Building on this taxonomy, we make three key contributions. First, we revisit recent findings suggesting that many removal methods may functionally behave like suppression, and explore whether true removal is necessary or achievable. Second, we survey existing evaluation strategies, identify limitations in current metrics and benchmarks, and suggest directions for developing more reliable and intention-aligned evaluations. Third, we highlight practical challenges--such as scalability and support for sequential unlearning--that currently hinder the broader deployment of unlearning methods. In summary, this work offers a comprehensive framework for understanding and advancing unlearning in generative AI, aiming to support future research and guide policy decisions around data removal and privacy. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:30:39 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Ren",
"Jie",
""
],
[
"Xing",
"Yue",
""
],
[
"Cui",
"Yingqian",
""
],
[
"Aggarwal",
"Charu C.",
""
],
[
"Liu",
"Hui",
""
]
] |
2506.09228 | Sita Kandel | Sita Kandel, Godfrey Gumbs, Teresa lee and Oleg L. Berman | Tuning excitons and superfluidity of dipolar excitons in the double layers of kagome lattice by applying circularly polarized irradiation | null | null | null | null | cond-mat.mes-hall | http://creativecommons.org/licenses/by/4.0/ | We present detailed calculations for several significant properties of the kagome lattice. We employ the Floquet-Magnus perturbation expansion to obtain the energy bands and the corresponding wave functions near the Dirac points for the kagome lattice in the presence of circularly or linearly polarized irradiation. In contrast with linearly polarized irradiation, a band gap is opened up near the Dirac points, between the valence and conduction bands in the presence of circularly polarized irradiation. We calculated the exciton binding energy, and the exciton energy for gapped kagome lattice as a function of the frequency and intensity of the irradiation. We compare the exciton binding energy and exciton energy in a monolayer with those in a double layer separated by an insulator to inhibit recombination. We predict that a phase transition in the kagome lattice from the semiconducting phase to the excitonic insulating phase can be induced by applying irradiation. We also examined the conditions for such a phase transition. We explore opportunities to tune exciton binding energy, the energy spectrum of collective excitations, the sound velocity and the critical temperature of the superfluidity by applying circularly polarized irradiation. We propose observation of Bose-Einstein condensation and superfluidity of quasi-two-dimensional dipolar excitons in two-layer kagome lattices in the presence of pumping by circularly polarized light. We have also analyzed the dependence of superfluid density $n_s$ and the temperature of the Kosterlitz-Thouless phase transition temperature on excitonic density n, the interlayer separation D and the parameters for circularly polarized light. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:33:49 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Kandel",
"Sita",
""
],
[
"Gumbs",
"Godfrey",
""
],
[
"lee",
"Teresa",
""
],
[
"Berman",
"Oleg L.",
""
]
] |
2506.09229 | Sungwon Hwang | Sungwon Hwang, Hyojin Jang, Kinam Kim, Minho Park, Jaegul choo | Cross-Frame Representation Alignment for Fine-Tuning Video Diffusion Models | 24 pages, 25 figures | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Fine-tuning Video Diffusion Models (VDMs) at the user level to generate videos that reflect specific attributes of training data presents notable challenges, yet remains underexplored despite its practical importance. Meanwhile, recent work such as Representation Alignment (REPA) has shown promise in improving the convergence and quality of DiT-based image diffusion models by aligning, or assimilating, its internal hidden states with external pretrained visual features, suggesting its potential for VDM fine-tuning. In this work, we first propose a straightforward adaptation of REPA for VDMs and empirically show that, while effective for convergence, it is suboptimal in preserving semantic consistency across frames. To address this limitation, we introduce Cross-frame Representation Alignment (CREPA), a novel regularization technique that aligns hidden states of a frame with external features from neighboring frames. Empirical evaluations on large-scale VDMs, including CogVideoX-5B and Hunyuan Video, demonstrate that CREPA improves both visual fidelity and cross-frame semantic coherence when fine-tuned with parameter-efficient methods such as LoRA. We further validate CREPA across diverse datasets with varying attributes, confirming its broad applicability. Project page: https://crepavideo.github.io | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:34:47 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Hwang",
"Sungwon",
""
],
[
"Jang",
"Hyojin",
""
],
[
"Kim",
"Kinam",
""
],
[
"Park",
"Minho",
""
],
[
"choo",
"Jaegul",
""
]
] |
2506.09230 | Juan Carlos Recio Abad Mr | Juan Carlos Recio Abad, Ruben Saborido and Francisco Chicano | Formal Methods Meets Readability: Auto-Documenting JML Java Code | null | null | null | null | cs.SE | http://creativecommons.org/licenses/by/4.0/ | This paper investigates whether formal specifications using Java Modeling Language (JML) can enhance the quality of Large Language Model (LLM)-generated Javadocs. While LLMs excel at producing documentation from code alone, we hypothesize that incorporating formally verified invariants yields more complete and accurate results. We present a systematic comparison of documentation generated from JML-annotated and non-annotated Java classes, evaluating quality through both automated metrics and expert analysis. Our findings demonstrate that JML significantly improves class-level documentation completeness, with more moderate gains at the method level. Formal specifications prove particularly effective in capturing complex class invariants and design contracts that are frequently overlooked in code-only documentation. A threshold effect emerges, where the benefits of JML become more pronounced for classes with richer sets of invariants. While JML enhances specification coverage, its impact on core descriptive quality is limited, suggesting that formal specifications primarily ensure comprehensive coverage rather than fundamentally altering implementation descriptions. These results offer actionable insights for software teams adopting formal methods in documentation workflows, highlighting scenarios where JML provides clear advantages. The study contributes to AI-assisted software documentation research by demonstrating how formal methods and LLMs can synergistically improve documentation quality. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:39:29 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Abad",
"Juan Carlos Recio",
""
],
[
"Saborido",
"Ruben",
""
],
[
"Chicano",
"Francisco",
""
]
] |
2506.09231 | Saba Tabatabaee | Saba Tabatabaee, Suzanne Boyce, Liran Oren, Mark Tiede, Carol Espy-Wilson | Enhancing Acoustic-to-Articulatory Speech Inversion by Incorporating Nasality | Accepted to be presented at Interspeech 2025 | null | null | null | eess.AS | http://creativecommons.org/licenses/by/4.0/ | Speech is produced through the coordination of vocal tract constricting organs: lips, tongue, velum, and glottis. Previous works developed Speech Inversion (SI) systems to recover acoustic-to-articulatory mappings for lip and tongue constrictions, called oral tract variables (TVs), which were later enhanced by including source information (periodic and aperiodic energies, and F0 frequency) as proxies for glottal control. Comparison of the nasometric measures with high-speed nasopharyngoscopy showed that nasalance can serve as ground truth, and that an SI system trained with it reliably recovers velum movement patterns for American English speakers. Here, two SI training approaches are compared: baseline models that estimate oral TVs and nasalance independently, and a synergistic model that combines oral TVs and source features with nasalance. The synergistic model shows relative improvements of 5% in oral TVs estimation and 9% in nasalance estimation compared to the baseline models. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:39:33 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Tabatabaee",
"Saba",
""
],
[
"Boyce",
"Suzanne",
""
],
[
"Oren",
"Liran",
""
],
[
"Tiede",
"Mark",
""
],
[
"Espy-Wilson",
"Carol",
""
]
] |
2506.09232 | Igor Zelenko | Nicklas Day, Boris Doubrov, and Igor Zelenko | Symplectification of Rank 2 Distributions, Normal Cartan Connections, and Cartan Prolongations | 22 pages | null | null | null | math.DG | http://creativecommons.org/licenses/by/4.0/ | We study the Doubrov--Zelenko symplectification procedure for rank~2 distributions of maximal class -- originally motivated by optimal control theory -- through the lens of Tanaka--Morimoto theory for normal Cartan connections. In this way, for ambient manifolds of dimension $n \geq 5$, we prove the existence of the normal Cartan connection associated with the symplectified distribution. Furthermore, we show that this symplectification can be interpreted as the $(n-4)$th iterated Cartan prolongation at a generic point. This interpretation naturally leads to two questions for an arbitrary rank $2$ distribution of maximal class: (1) Is the $(n-4)$th iterated Cartan prolongation the minimal iteration where the Tanaka symbols become unified at generic points? (2) Is the $(n-4)$th iterated Cartan prolongation the minimal iteration admitting a normal Cartan connection via Tanaka--Morimoto theory? Our main results demonstrate that: (a) For $n > 5$, the answer to the second question is positive (in contrast to the classical $n = 5$ case from $G_2$-parabolic geometries); (b) For $n \geq 5$, the answer to the first question is negative: unification occurs already at the $(n-5)$th iterated Cartan prolongation. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:39:54 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Day",
"Nicklas",
""
],
[
"Doubrov",
"Boris",
""
],
[
"Zelenko",
"Igor",
""
]
] |
2506.09233 | Xianzhu Tang | Qile Zhang and Yanzeng Zhang and Xian-Zhu Tang | Excitation of whistler and slow-X waves by runaway electrons in a collisional plasma | 15 pages, 15 figures | null | null | LA-UR-25-25500 | physics.plasm-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Runaway electrons are known to provide robust ideal or collisionless
kinetic drive for plasma wave instabilities in both the whistler and
slow-X branches, via the anomalous Doppler-shifted cyclotron
resonances. In a cold and dense post-thermal-quench plasma,
collisional damping of the plasma waves can be competitive with the
collisionless drive. Previous studies have found that for its higher
wavelength and frequency, slow-X waves suffer stronger collisional
damping than the whistlers, while the ideal growth rate of slow-X
modes is higher. Here we study runaway avalanche distributions that maintain the same eigen distribution and increase only in magnitude over time. The distributions are computed from the
relativistic Fokker-Planck-Boltzmann solver, upon which a linear
dispersion analysis is performed to search for the most unstable or
least damped slow-X and whistler modes. Taking into account the
effect of plasma density, plasma temperature, and effective charge
number, we find that the slow-X modes tend to be excited before the
whistlers in a runaway current ramp-up. Furthermore, even when the
runaway current density is sufficiently high that both branches are
excited, the most unstable slow-X mode has much higher growth rate
than the most unstable whistler mode. The qualitative and
quantitative trends uncovered in current study indicate that even
though past experiments and modeling efforts have concentrated on whistler modes, there's a compelling case that slow-X modes should also be a key area of focus. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:41:05 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Zhang",
"Qile",
""
],
[
"Zhang",
"Yanzeng",
""
],
[
"Tang",
"Xian-Zhu",
""
]
] |
2506.09234 | Kaiwen Dong | Kaiwen Dong, Padmaja Jonnalagedda, Xiang Gao, Ayan Acharya, Maria Kissa, Mauricio Flores, Nitesh V. Chawla, and Kamalika Das | Transaction Categorization with Relational Deep Learning in QuickBooks | Accepted to ECML-PKDD 2025 | null | null | null | cs.CE | http://creativecommons.org/licenses/by/4.0/ | Automatic transaction categorization is crucial for enhancing the customer experience in QuickBooks by providing accurate accounting and bookkeeping. The distinct challenges in this domain stem from the unique formatting of transaction descriptions, the wide variety of transaction categories, and the vast scale of the data involved. Furthermore, organizing transaction data in a relational database creates difficulties in developing a unified model that covers the entire database. In this work, we develop a novel graph-based model, named Rel-Cat, which is built directly over the relational database. We introduce a new formulation of transaction categorization as a link prediction task within this graph structure. By integrating techniques from natural language processing and graph machine learning, our model not only outperforms the existing production model in QuickBooks but also scales effectively to a growing customer base with a simpler, more effective architecture without compromising on accuracy. This design also helps tackle a key challenge of the cold start problem by adapting to minimal data. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:42:41 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Dong",
"Kaiwen",
""
],
[
"Jonnalagedda",
"Padmaja",
""
],
[
"Gao",
"Xiang",
""
],
[
"Acharya",
"Ayan",
""
],
[
"Kissa",
"Maria",
""
],
[
"Flores",
"Mauricio",
""
],
[
"Chawla",
"Nitesh V.",
""
],
[
"Das",
"Kamalika",
""
]
] |
2506.09235 | Caitlin Kengle | Caitlin S. Kengle, Noah Schnitzer, Elizabeth A. Peterson, Chunyu Guo, Ling Zhang, Matthew S. Cook, Jian-Xin Zhu, Sean M. Thomas, Philip J. W. Moll, Filip Ronning, Priscila F.S. Rosa | Putative excitonic insulating state in narrow-gap semiconductor La$_3$Cd$_2$As$_6$ | LA-UR-25-25525 | null | null | null | cond-mat.str-el | http://creativecommons.org/licenses/by/4.0/ | Excitonic insulators are electronically-driven phases of matter characterized by the spontaneous condensation of electron-hole pairs. Here we show that La$_3$Cd$_2$As$_6$ undergoes a transition at $T_{0}=278$ K to a highly insulating state with no accompanying structural transition. We observe quasi-two-dimensional electrical transport and charge fluctuations consistent with an electronic transition enabled by enhanced Coulomb interactions. Density functional theory calculations are unable to replicate the insulating ground state. Our results support the opening of a gap by excitonic effects at $T_{0}$, placing La$_3$Cd$_2$As$_6$ as a rare example of a bulk excitonic insulator. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:43:17 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Kengle",
"Caitlin S.",
""
],
[
"Schnitzer",
"Noah",
""
],
[
"Peterson",
"Elizabeth A.",
""
],
[
"Guo",
"Chunyu",
""
],
[
"Zhang",
"Ling",
""
],
[
"Cook",
"Matthew S.",
""
],
[
"Zhu",
"Jian-Xin",
""
],
[
"Thomas",
"Sean M.",
""
],
[
"Moll",
"Philip J. W.",
""
],
[
"Ronning",
"Filip",
""
],
[
"Rosa",
"Priscila F. S.",
""
]
] |
2506.09236 | Regis Kopper | Erin Argo, Tanim Ahmed, Sarah Gable, Callie Hampton, Jeronimo Grandi, Regis Kopper | Augmented Reality User Interfaces for First Responders: A Scoping Literature Review | 19 pages, 4 figures, 8 tables | null | null | null | cs.HC | http://creativecommons.org/licenses/by-sa/4.0/ | During the past decade, there has been a significant increase in research focused on integrating AR User Interfaces into public safety applications, particularly for first responders in the domains of Emergency Medical Services, Firefighting, and Law Enforcement. This paper presents the results of a scoping review involving the application of AR user interfaces in the public safety domain and applies an established systematic review methodology to provide a comprehensive analysis of the current research landscape, identifying key trends, challenges, and gaps in the literature. This review includes peer-reviewed publications indexed by the major scientific databases up to April 2025. A basic keyword search retrieved 1,751 papers, of which 90 were deemed relevant for this review. An in-depth analysis of the literature allowed the development of a faceted taxonomy that categorizes AR user interfaces for public safety. This classification lays a solid foundation for future research, while also highlighting key design considerations, challenges, and gaps in the literature. This review serves as a valuable resource for researchers and developers, offering insights that can drive further advances in the field. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:44:15 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Argo",
"Erin",
""
],
[
"Ahmed",
"Tanim",
""
],
[
"Gable",
"Sarah",
""
],
[
"Hampton",
"Callie",
""
],
[
"Grandi",
"Jeronimo",
""
],
[
"Kopper",
"Regis",
""
]
] |
2506.09237 | Mojtaba Nafez | Mojtaba Nafez, Amirhossein Koochakian, Arad Maleki, Jafar Habibi, Mohammad Hossein Rohban | PatchGuard: Adversarially Robust Anomaly Detection and Localization through Vision Transformers and Pseudo Anomalies | Accepted to the Conference on Computer Vision and Pattern Recognition (CVPR) 2025 | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Anomaly Detection (AD) and Anomaly Localization (AL) are crucial in fields that demand high reliability, such as medical imaging and industrial monitoring. However, current AD and AL approaches are often susceptible to adversarial attacks due to limitations in training data, which typically include only normal, unlabeled samples. This study introduces PatchGuard, an adversarially robust AD and AL method that incorporates pseudo anomalies with localization masks within a Vision Transformer (ViT)-based architecture to address these vulnerabilities. We begin by examining the essential properties of pseudo anomalies, and follow it by providing theoretical insights into the attention mechanisms required to enhance the adversarial robustness of AD and AL systems. We then present our approach, which leverages Foreground-Aware Pseudo-Anomalies to overcome the deficiencies of previous anomaly-aware methods. Our method incorporates these crafted pseudo-anomaly samples into a ViT-based framework, with adversarial training guided by a novel loss function designed to improve model robustness, as supported by our theoretical analysis. Experimental results on well-established industrial and medical datasets demonstrate that PatchGuard significantly outperforms previous methods in adversarial settings, achieving performance gains of $53.2\%$ in AD and $68.5\%$ in AL, while also maintaining competitive accuracy in non-adversarial settings. The code repository is available at https://github.com/rohban-lab/PatchGuard . | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:45:54 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Nafez",
"Mojtaba",
""
],
[
"Koochakian",
"Amirhossein",
""
],
[
"Maleki",
"Arad",
""
],
[
"Habibi",
"Jafar",
""
],
[
"Rohban",
"Mohammad Hossein",
""
]
] |
2506.09238 | Chuyue Fang | David Caratelli, Nathaniel Craig, Chuyue Fang, Jessica N. Howard | Optimal Transport for $e/\pi^0$ Particle Classification in LArTPC Neutrino Experiments | 19 pages, 13 figures, 2 tables | null | null | null | hep-ex physics.ins-det | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The efficient classification of electromagnetic activity from $\pi^0$ and electrons remains an open problem in the reconstruction of neutrino interactions in Liquid Argon Time Projection Chamber (LArTPC) detectors. We address this problem using the mathematical framework of Optimal Transport (OT), which has been successfully employed for event classification in other HEP contexts and is ideally suited to the high-resolution calorimetry of LArTPCs. Using a publicly available simulated dataset from the MicroBooNE collaboration, we show that OT methods achieve state-of-the-art reconstruction performance in $e/\pi^0$ classification. The success of this first application indicates the broader promise of OT methods for LArTPC-based neutrino experiments. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:49:59 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Caratelli",
"David",
""
],
[
"Craig",
"Nathaniel",
""
],
[
"Fang",
"Chuyue",
""
],
[
"Howard",
"Jessica N.",
""
]
] |
2506.09239 | Cheuk Ting Li | Cheuk Ting Li and Jianguo Zhao | Rejection-Sampled Linear Codes for Lossy Compression and Channel Simulation | 12 pages, 5 figures | null | null | null | cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We show that a linear code combined with rejection sampling can give a capacity-achieving scheme for simulating channels with additive noises with exchangeable distributions. Hence, it can be used in lossy source coding to achieve the rate-distortion function. Interestingly, unlike conventional linear covering codes for lossy compression which concerns the trade-off between the rate and the covering radius, our construction only requires the linear code to have a large distance (not a large covering radius), and is not sensitive to the rate of the linear code. Experiments reveal that our construction can outperform conventional covering codes for lossy source coding with Hamming distortion for a certain range of distortion levels, and performs well even when the blocklength is small (e.g., 24). | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:52:19 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Li",
"Cheuk Ting",
""
],
[
"Zhao",
"Jianguo",
""
]
] |
2506.09240 | Amanda Wang | Amanda Wang, Nick Pant, Woncheol Lee, Jie-Cheng Chen, Feliciano Giustino, Emmanouil Kioupakis | Electron mobility in AlN from first principles | 18 pages, 2 figures in main text, 2 in supplementary | null | null | null | cond-mat.mtrl-sci | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Aluminum nitride is a promising ultra-wide band gap semiconductor for optoelectronics and power electronics. However, its practical applications have been limited by challenges with doping and achieving high electrical conductivity. Recent advances in crystal quality and defect control have led to improvements in experimentally measured mobilities. In this work, we apply first-principles calculations to determine the upper limits of the electron mobility in AlN as a function of temperature, doping, and crystallographic orientation. We account for the combined effects of electron scattering by phonons and ionized impurity to model doped systems, and examine both full and partial ionization conditions. Our results show that the piezoelectric interaction from the long-range component of the acoustic modes is the dominant source of electron-phonon scattering at room temperature. Ionized-impurity scattering starts to dominate scattering at dopant concentrations above $10^{16}$ cm$^{-3}$, reducing the mobility by more than an order of magnitude in the high doping regime. Our calculated Hall mobility values are in good agreement with experimental data for samples with comparable dopant concentrations. We also find that electron mobilities as high as $956$ cm$^2$/V$\cdot$s could be achievable at lower dopant concentrations. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:53:15 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Wang",
"Amanda",
""
],
[
"Pant",
"Nick",
""
],
[
"Lee",
"Woncheol",
""
],
[
"Chen",
"Jie-Cheng",
""
],
[
"Giustino",
"Feliciano",
""
],
[
"Kioupakis",
"Emmanouil",
""
]
] |
2506.09241 | Majid Akbari | Majid Akbari | K-Exit for OD-Characterization of the finite Groups | null | null | null | null | math.GR | http://creativecommons.org/licenses/by/4.0/ | Suppose G is a finite group, and K is a normal solvable subgroup of G. In this paper, we present a
powerful method which is called K-Exit. Using this method, we find some members of {\pi}(G) that
are not members of {\pi}(K) and for this purpose, we introduce the tables of K-Exit. This method
can be applied to OD-Characterization of the finite Groups | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:55:13 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Akbari",
"Majid",
""
]
] |
2506.09242 | Jonas Latt | Jonas Latt and Christophe Coreixas | Multi-GPU Acceleration of PALABOS Fluid Solver using C++ Standard Parallelism | null | null | null | null | cs.DC | http://creativecommons.org/licenses/by/4.0/ | This article presents the principles, software architecture, and performance analysis of the GPU port of the lattice Boltzmann software library Palabos (J. Latt et al., "Palabos: Parallel lattice Boltzmann solver", Comput. Math. Appl. 81, 334-350, (2021)). A hybrid CPU-GPU execution model is adopted, in which numerical components are selectively assigned to either the CPU or the GPU, depending on considerations of performance or convenience. This design enables a progressive porting strategy, allowing most features of the original CPU-based codebase to be gradually and seamlessly adapted to GPU execution. The new architecture builds upon two complementary paradigms: a classical object-oriented structure for CPU execution, and a data-oriented counterpart for GPUs, which reproduces the modularity of the original code while eliminating object-oriented overhead detrimental to GPU performance. Central to this approach is the use of modern C++, including standard parallel algorithms and template metaprogramming techniques, which permit the generation of hardware-agnostic computational kernels. This facilitates the development of user-defined, GPU-accelerated components such as collision operators or boundary conditions, while preserving compatibility with the existing codebase and avoiding the need for external libraries or non-standard language extensions. The correctness and performance of the GPU-enabled Palabos are demonstrated through a series of three-dimensional multiphysics benchmarks, including the laminar-turbulent transition in a Taylor-Green vortex, lid-driven cavity flow, and pore-scale flow in Berea sandstone. Despite the high-level abstraction of the implementation, the single-GPU performance is similar to CUDA-native solvers, and multi-GPU tests exhibit good weak and strong scaling across all test cases. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:55:28 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Latt",
"Jonas",
""
],
[
"Coreixas",
"Christophe",
""
]
] |
2506.09243 | Andrey Matsko | Andrey Matsko, Abdelkrim El Amili, Lute Maleki | Kerr nonlinearity, self-injection locking and correlation in a microresonator | null | null | null | null | physics.optics | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Production of entangled photon pairs is important in secure communication systems, quantum computing, and fundamental physics experiments. Achieving efficient generation of such photon pairs with low-loss parametric oscillators is a key objective in advancing integrated quantum technologies. However, spatially separating the generated photons while preserving their entanglement represents a significant technical challenge. In this work, we demonstrate nonlinear generation of correlated optical harmonics based on non-degenerate four-wave mixing with an optimally pumped optical microcavity with Kerr nonlinearity. The phase matching of the process is achieved with self-injection locked lasers producing parametric oscillation while locked to two different modes of the microresonator. This condition is reminiscent of slow-light technique developed for coherent atomic systems. The experimental design, utilizing counterpropagating light from two self-injection locked lasers, also effectively addresses the challenge of spatial separation of the generated harmonics. We validate the theoretical predictions using two self-injection locked semiconductor lasers integrated with a crystalline whispering gallery mode resonator with optimized spectral structure. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:55:29 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Matsko",
"Andrey",
""
],
[
"Amili",
"Abdelkrim El",
""
],
[
"Maleki",
"Lute",
""
]
] |
2506.09244 | Damir Kinzebulatov | D.Kinzebulatov and R.Vafadar | SDEs with critical general distributional drifts: sharp solvability and blow ups | null | null | null | null | math.PR math-ph math.AP math.MP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We establish weak well-posedness for SDEs having discontinuous diffusion coefficients and general distributional drifts that may introduce blow up effects. Our drifts satisfy minimal assumptions, i.e.\,we assume only that the Cauchy problem for the Kolmogorov backward equation is well-posed in the standard Hilbert triple $W^{1,2} \hookrightarrow L^2 \hookrightarrow W^{-1,2}$. By a result of Mazya and Verbitsky, these assumptions are precisely those drifts that can be represented as the sum of a form-bounded component (encompassing, for example, Morrey or Chang-Wilson-Wolff drifts) and a divergence-free distributional component in the ${\rm BMO}^{-1}$ space of Koch and Tataru.
We apply our results to finite particle systems with strong attracting interactions immersed in a turbulent flow. This includes particle systems of Keller-Segel type. Crucially, in dimensions $d \geq 3$, we cover almost the entire admissible range of attraction strengths, reaching nearly to the blow-up threshold.
As a further application of our results for SDEs and of the theory of Bessel processes, we obtain an improved upper bound on the constant in the many-particle Hardy inequality. Consequently, the lower bound previously derived by Hoffmann-Ostenhof, Hoffmann-Ostenhof, Laptev, and Tidblom is shown to be close to optimal. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 20:56:59 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Kinzebulatov",
"D.",
""
],
[
"Vafadar",
"R.",
""
]
] |
2506.09245 | Aresh Dadlani | Muthukrishnan Senthilkumar, Aresh Dadlani, Hina Tabassum | Age of Information in Unreliable Tandem Queues | null | null | null | null | cs.NI | http://creativecommons.org/licenses/by/4.0/ | Stringent demands for timely information delivery, driven by the widespread adoption of real-time applications and the Internet of Things, have established the age of information (AoI) as a critical metric for quantifying data freshness. Existing AoI models often assume multi-hop communication networks with fully reliable nodes, which may not accurately capture scenarios involving node transmission failures. This paper presents an analytical framework for two configurations of tandem queue systems, where status updates generated by a single sensor are relayed to a destination monitor through unreliable intermediate nodes. Using the probability generating function, we first derive the sojourn time distribution for an infinite-buffer M/M/1 tandem system with two unreliable nodes. We then extend our analysis to an M/G/1 tandem system with an arbitrary number of unreliable nodes, employing the supplementary variable technique while assuming that only the first node has an infinite buffer. Numerical results demonstrate the impact of key system parameters on the average AoI in unreliable tandem queues with Markovian and non-Markovian service times. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 21:02:32 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Senthilkumar",
"Muthukrishnan",
""
],
[
"Dadlani",
"Aresh",
""
],
[
"Tabassum",
"Hina",
""
]
] |
2506.09246 | Catharina Stroppel | Thorsten Heidersdorf and Jonas Nehme and Catharina Stroppel | Koszulity for semi-infinite highest weight categories | null | null | null | MPIM-Bonn-2025 | math.RT | http://creativecommons.org/licenses/by-nc-sa/4.0/ | We show that any upper finite or essentially finite highest weight category where the standard objects have linear projective resolutions and the costandard objects have linear injective resolutions is Koszul. This extends the result of Agoston, Dlab, and Lukacs to the case of infinite highest weight categories. We apply this result to Khovanov algebras and representations of classical Deligne categories and show that these are Koszul. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 21:03:52 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Heidersdorf",
"Thorsten",
""
],
[
"Nehme",
"Jonas",
""
],
[
"Stroppel",
"Catharina",
""
]
] |
2506.09247 | Karl L\"owenmark Dr. | Karl L\"owenmark, Daniel Str\"ombergsson, Chang Liu, Marcus Liwicki, Fredrik Sandin | Agent-based Condition Monitoring Assistance with Multimodal Industrial Database Retrieval Augmented Generation | null | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Condition monitoring (CM) plays a crucial role in ensuring reliability and efficiency in the process industry. Although computerised maintenance systems effectively detect and classify faults, tasks like fault severity estimation, and maintenance decisions still largely depend on human expert analysis. The analysis and decision making automatically performed by current systems typically exhibit considerable uncertainty and high false alarm rates, leading to increased workload and reduced efficiency.
This work integrates large language model (LLM)-based reasoning agents with CM workflows to address analyst and industry needs, namely reducing false alarms, enhancing fault severity estimation, improving decision support, and offering explainable interfaces. We propose MindRAG, a modular framework combining multimodal retrieval-augmented generation (RAG) with novel vector store structures designed specifically for CM data. The framework leverages existing annotations and maintenance work orders as surrogates for labels in a supervised learning protocol, addressing the common challenge of training predictive models on unlabelled and noisy real-world datasets.
The primary contributions include: (1) an approach for structuring industry CM data into a semi-structured multimodal vector store compatible with LLM-driven workflows; (2) developing multimodal RAG techniques tailored for CM data; (3) developing practical reasoning agents capable of addressing real-world CM queries; and (4) presenting an experimental framework for integrating and evaluating such agents in realistic industrial scenarios. Preliminary results, evaluated with the help of an experienced analyst, indicate that MindRAG provide meaningful decision support for more efficient management of alarms, thereby improving the interpretability of CM systems. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 21:04:18 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Löwenmark",
"Karl",
""
],
[
"Strömbergsson",
"Daniel",
""
],
[
"Liu",
"Chang",
""
],
[
"Liwicki",
"Marcus",
""
],
[
"Sandin",
"Fredrik",
""
]
] |
2506.09248 | Vinicius F. Dal Poggetto | V. F. Dal Poggetto and M. Miniaci | Ultra-thin normal-shear-coupled metabarrier for low-frequency underwater sound insulation | null | null | null | null | physics.app-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Underwater noise pollution caused by anthropogenic activities, such as offshore wind farms, significantly affects marine life, hindering the intra- and inter-specific interactions of many aquatic species. A common strategy to mitigate these effects is to enclose the noise source within a physical barrier to achieve acceptable noise levels in the surrounding region. Underwater barriers typically achieve noise reduction through sound absorption based on locally resonant systems (e.g., foam and bubble elements) or sound reflecting systems (e.g., air bubble curtains). Although locally resonant-based solutions can yield significant noise attenuation levels, their performance is usually narrow-band. On the other hand, sound-reflecting systems require wider dimensions, presenting poor performance in the low-frequency range. Thus, significant low-frequency underwater noise attenuation remains an open issue, especially when considering thin structures that perform over a broad frequency range. In this work, we present the design of thin metamaterial-based acoustic barriers whose underwater noise attenuation is owed to tailored anisotropic material properties. A topology optimization approach is used to obtain a unit cell that maximizes the coupling between normal stresses and shear strains (and vice-versa). The resulting metabarrier presents a sub-wavelength thickness-to-wavelength ratio in the low-frequency range (circa 1/70 below 1 kHz) and also high sound transmission loss values at higher frequencies (almost 100 dB above 2 kHz). We also investigate the effects of increased hydrostatic pressure, presenting structural modifications that enable real-world applications. The results presented in this work indicate an efficient metamaterial-based solution for the mitigation of underwater noise, suggesting a fruitful direction for the exploitation of anisotropy in acoustic insulation applications. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 21:16:19 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Poggetto",
"V. F. Dal",
""
],
[
"Miniaci",
"M.",
""
]
] |
2506.09249 | Sebastian Halbig | Sebastian Halbig and Ulrich Kr\"ahmer | A non-semisimple Kitaev lattice model | comments very welcome | null | null | null | math.QA | http://creativecommons.org/licenses/by-nc-nd/4.0/ | The construction of the topologically protected code space of Kitaev's model for fault-tolerant quantum computation is extended from complex semisimple to arbitrary finite-dimensional Hopf algebras admitting pairs in involution. One input of the model are ribbon graphs, that is, the combinatorial data of cellular decompositions of oriented closed surfaces. The other input are certain Hopf bimodules that are closely related to the coefficients in Hopf-cyclic homology. As in previous generalisations of the Kitaev model, a Yetter-Drinfeld module is constructed and shown to be a topological invariant of the surface with boundary that is obtained by "thickening" the graph. The generalisation of the protected space is defined using bitensor products of modules-comodules. Provided that the Hopf bimodule coefficients correspond to pairs in involution, this is shown to depend only on the genus of the graph. As examples, group algebras of finite groups and bosonisations of Nichols algebras are considered. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 21:16:24 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Halbig",
"Sebastian",
""
],
[
"Krähmer",
"Ulrich",
""
]
] |
2506.09250 | Alex Lawsen | C. Opus and A. Lawsen | Comment on The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity | Comment on: arXiv:2506.06941 | null | null | null | cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Shojaee et al. (2025) report that Large Reasoning Models (LRMs) exhibit "accuracy collapse" on planning puzzles beyond certain complexity thresholds. We demonstrate that their findings primarily reflect experimental design limitations rather than fundamental reasoning failures. Our analysis reveals three critical issues: (1) Tower of Hanoi experiments systematically exceed model output token limits at reported failure points, with models explicitly acknowledging these constraints in their outputs; (2) The authors' automated evaluation framework fails to distinguish between reasoning failures and practical constraints, leading to misclassification of model capabilities; (3) Most concerningly, their River Crossing benchmarks include mathematically impossible instances for N > 5 due to insufficient boat capacity, yet models are scored as failures for not solving these unsolvable problems. When we control for these experimental artifacts, by requesting generating functions instead of exhaustive move lists, preliminary experiments across multiple models indicate high accuracy on Tower of Hanoi instances previously reported as complete failures. These findings highlight the importance of careful experimental design when evaluating AI reasoning capabilities. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 21:16:53 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Opus",
"C.",
""
],
[
"Lawsen",
"A.",
""
]
] |
2506.09251 | Ziyang Cai | Ziyang Cai, Nayoung Lee, Avi Schwarzschild, Samet Oymak, Dimitris Papailiopoulos | Extrapolation by Association: Length Generalization Transfer in Transformers | 23 pages, 20 figures | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Transformer language models have demonstrated impressive generalization capabilities in natural language domains, yet we lack a fine-grained understanding of how such generalization arises. In this paper, we investigate length generalization--the ability to extrapolate from shorter to longer inputs--through the lens of \textit{task association}. We find that length generalization can be \textit{transferred} across related tasks. That is, training a model with a longer and related auxiliary task can lead it to generalize to unseen and longer inputs from some other target task. We demonstrate this length generalization transfer across diverse algorithmic tasks, including arithmetic operations, string transformations, and maze navigation. Our results show that transformer models can inherit generalization capabilities from similar tasks when trained jointly. Moreover, we observe similar transfer effects in pretrained language models, suggesting that pretraining equips models with reusable computational scaffolding that facilitates extrapolation in downstream settings. Finally, we provide initial mechanistic evidence that length generalization transfer correlates with the re-use of the same attention heads between the tasks. Together, our findings deepen our understanding of how transformers generalize to out-of-distribution inputs and highlight the compositional reuse of inductive structure across tasks. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 21:22:51 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Cai",
"Ziyang",
""
],
[
"Lee",
"Nayoung",
""
],
[
"Schwarzschild",
"Avi",
""
],
[
"Oymak",
"Samet",
""
],
[
"Papailiopoulos",
"Dimitris",
""
]
] |
2506.09252 | Makumburage Don Hashan C. Peiris | M.D. Hashan C. Peiris, Michael Woodcox, Diana Liepinya, Robert Shephard, Hao Liu, Manuel Smeu | Aluminum oxide coatings on Co-rich cathodes and interactions with organic electrolyte | Submitted for NIST Internal Review | null | null | null | cond-mat.mtrl-sci physics.chem-ph physics.comp-ph | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Lithium-ion batteries (LIBs) have become essential in modern energy storage; however, their performance is often limited by the stability and efficiency of their components, particularly the cathode and electrolyte. Transition metal layered oxide cathodes, a popular choice for lithium-ion batteries (LIBs), suffer from several degradation mechanisms, including capacity fading, reactions with the electrolyte, unstable cathode-electrolyte interfaces, and lattice breakdown during cycling. In recent years, oxide coating, such as alumina, has emerged as a promising strategy to enhance the durability of cathodes by forming a protective layer that mitigates detrimental reactions and improves the stability of the cathode electrolyte interphase (CEI). This study employs ab initio molecular dynamics (AIMD) simulations to investigate the chemical and mechanical behavior of LiCoO2 cathodes with and without aluminum oxide coatings in contact with an organic electrolyte. We examine the interactions between electrolyte molecules with both bare and coated cathode surfaces, focusing on the decomposition of ethylene carbonate (EC) and dimethyl carbonate (DMC), the formation of oxygen species, and solvation dynamics, and evaluate the mechanical robustness of the cathode-coating interface using calculations of axial strain and cleavage energy. Our findings reveal that alumina coatings effectively reduce electrolyte degradation and stabilize the cathode structure, particularly under high-charge states. The coating's thickness and structural orientation are crucial in enhancing mechanical strength and minimizing detrimental reactions at the cathode-electrolyte interface. These insights contribute to the development of more durable LIBs by optimizing the interface chemistry and mechanical properties, providing a pathway toward higher energy densities and longer cycle life. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 21:25:54 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Peiris",
"M. D. Hashan C.",
""
],
[
"Woodcox",
"Michael",
""
],
[
"Liepinya",
"Diana",
""
],
[
"Shephard",
"Robert",
""
],
[
"Liu",
"Hao",
""
],
[
"Smeu",
"Manuel",
""
]
] |
2506.09253 | Grant Kirchhoff | Grant J. Kirchhoff, Matthew Hayman, Willem J. Marais, Jeffrey P. Thayer, Rory A. Barton-Grimley | Development of a Photon-Counting Deadtime Noise Model that Extends Dynamic Range and Resolution in Atmospheric Lidar | null | Applied Optics 64 (2025) 4568-4581 | 10.1364/AO.543305 | null | eess.SP physics.ins-det | http://creativecommons.org/licenses/by/4.0/ | This work derives and validates a noise model that encapsulates deadtime of non-paralyzable detectors with random photon arrivals to enable advanced processing, like maximum-likelihood estimation, of high resolution atmospheric lidar profiles while accounting for deadtime bias. This estimator was validated across a wide dynamic range at high resolution (4 millimeters in range, 17 milliseconds in time). Experiments demonstrate that the noise model outperforms the current state-of-the-art for very short time-of-flight (2 nanoseconds) and extended targets (1 microsecond). The proposed noise model also produces accurate deadtime correction for very short integration times. This work sets the foundation for further study into accurate retrievals of high flux and dynamic atmospheric features, e.g., clouds and aerosol layers. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 21:27:14 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Kirchhoff",
"Grant J.",
""
],
[
"Hayman",
"Matthew",
""
],
[
"Marais",
"Willem J.",
""
],
[
"Thayer",
"Jeffrey P.",
""
],
[
"Barton-Grimley",
"Rory A.",
""
]
] |
2506.09255 | Saeed Hashemi | Saeed Hashemi, Genchang Peng, Mehrdad Nourani, Omar Nofal, Jay Harvey | AI-Driven SEEG Channel Ranking for Epileptogenic Zone Localization | Accepted to be presented at the 47th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC 2025). This version is submitted to arXiv prior to final IEEE formatting and publication | null | null | null | eess.SP cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Stereo-electroencephalography (SEEG) is an invasive technique to implant depth electrodes and collect data for pre-surgery evaluation. Visual inspection of signals recorded from hundreds of channels is time consuming and inefficient. We propose a machine learning approach to rank the impactful channels by incorporating clinician's selection and computational finding. A classification model using XGBoost is trained to learn the discriminative features of each channel during ictal periods. Then, the SHapley Additive exPlanations (SHAP) scoring is utilized to rank SEEG channels based on their contribution to seizures. A channel extension strategy is also incorporated to expand the search space and identify suspicious epileptogenic zones beyond those selected by clinicians. For validation, SEEG data for five patients were analyzed showing promising results in terms of accuracy, consistency, and explainability. | [
{
"version": "v1",
"created": "Fri, 23 May 2025 19:09:16 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Hashemi",
"Saeed",
""
],
[
"Peng",
"Genchang",
""
],
[
"Nourani",
"Mehrdad",
""
],
[
"Nofal",
"Omar",
""
],
[
"Harvey",
"Jay",
""
]
] |
2506.09256 | Logan Brabson | Logan M. Brabson, Andrew J. Medford, David S. Sholl | Comparing classical and machine learning force fields for modeling deformation of solid sorbents relevant for direct air capture | null | null | null | null | cond-mat.mtrl-sci physics.chem-ph | http://creativecommons.org/licenses/by/4.0/ | Direct air capture (DAC) with solid sorbents such as metal-organic frameworks (MOFs) is a promising approach for negative carbon emissions. Computational materials screening can help identify promising materials from the vast chemical space of potential sorbents. Experiments have shown that MOF framework flexibility and deformation induced by adsorbate molecules can drastically affect adsorption properties such as capacity and selectivity. Force field (FF) models are commonly used as surrogates for more accurate density functional theory (DFT) calculations when modeling sorbents, but most studies using FFs for MOFs assume framework rigidity to simplify calculations. Although flexible FFs for MOFs have been parameterized for specific materials, the generality of FFs for reliably modeling adsorbate-induced deformation to near-DFT accuracy has not been established. This work benchmarks the efficacy of several general FFs in describing adsorbate-induced deformation for DAC against DFT. Specifically, we compare a common classical FF (UFF4MOF) with several machine learning (ML) FFs: M3GNet, CHGNet, MACE-MP-0, MACE-MPA-0, eSEN, and the Equiformer V2 model developed from the recent Open DAC 2023 dataset. Our results show that current classical methods are insufficient for describing framework deformation, especially in cases of interest for DAC where strong interactions exist between adsorbed molecules and MOF frameworks. The emerging ML methods we tested -- particularly CHGNet, MACE-MP-0, and Equiformer V2 -- appear to be more promising than the classical FF for emulating the deformation behavior described by DFT but fail to achieve the accuracy required for practical predictions. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 21:34:26 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Brabson",
"Logan M.",
""
],
[
"Medford",
"Andrew J.",
""
],
[
"Sholl",
"David S.",
""
]
] |
2506.09257 | Glenn Orton | Glenn S. Orton, Magnus Gustafsson, Leigh N. Fletcher, Michael T. Roman, and James A. Sinclair | Improved H2-He and H2-H2 Collision-Induced Absorption Models and Application to Outer-Planet Atmospheres | 30 pages (including an Appendix), 13 figures, 3 tables | null | null | null | astro-ph.EP physics.chem-ph | http://creativecommons.org/licenses/by/4.0/ | Using state-of-the-art ab initio interaction-induced dipole and potential-energy surfaces for hydrogen-helium (H2-He) pairs, we compute the rototranslational collision-induced absorption coefficient at 40-400 K for frequencies covering 0-4000 cm-1. The quantum mechanical scattering calculations account for the full anisotropic interaction potential, replacing the isotropic approximation. The absorption data are expected to be accurate with an uncertainty of 2% or better up to 2500 cm-1. The uncertainty is slightly higher at the highest frequencies where the rototranslational absorption is largely obscured by the rovibrational band. Our improved agreement with measurements at 200-800 cm-1 results from the improvement of the potential energy surface. The previously available rototranslational data set for H2-H2 pairs (Fletcher et al., Astrophys. J. Supp. 235, 24 (2018)) is also extended up to 4000 cm-1. In the rovibrational band previous isotropic potential calculations for H2-He (Gustafsson et al. J. Chem. Physics. 113, 3641 (2000)) and H2-H2 (Borysow, Icarus 92, 273 (1992)) have been extended to complement the rototranslational data set. The absorption coefficients are tabulated for ortho-to-para ratios from normal-H2 to pure para-H2, as well as equilibrium-H2, over 40-400 K. The effect of these updates are simulated for the cold atmosphere of Uranus and warmer atmosphere of Jupiter. They are equivalent to a brightness temperature difference of a fraction of a degree in the rototranslational region but up to 4 degrees in the rovibrational region. Our state-of-the-art modifications correct an otherwise +2% error in determining the He/H2 ratio in Uranus from its spectrum alone. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 21:35:22 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Orton",
"Glenn S.",
""
],
[
"Gustafsson",
"Magnus",
""
],
[
"Fletcher",
"Leigh N.",
""
],
[
"Roman",
"Michael T.",
""
],
[
"Sinclair",
"James A.",
""
]
] |
2506.09258 | Vaidotas Simkus | Vaidotas Simkus, Michael U. Gutmann | CFMI: Flow Matching for Missing Data Imputation | null | null | null | null | cs.LG stat.ML | http://creativecommons.org/licenses/by/4.0/ | We introduce conditional flow matching for imputation (CFMI), a new general-purpose method to impute missing data. The method combines continuous normalising flows, flow-matching, and shared conditional modelling to deal with intractabilities of traditional multiple imputation. Our comparison with nine classical and state-of-the-art imputation methods on 24 small to moderate-dimensional tabular data sets shows that CFMI matches or outperforms both traditional and modern techniques across a wide range of metrics. Applying the method to zero-shot imputation of time-series data, we find that it matches the accuracy of a related diffusion-based method while outperforming it in terms of computational efficiency. Overall, CFMI performs at least as well as traditional methods on lower-dimensional data while remaining scalable to high-dimensional settings, matching or exceeding the performance of other deep learning-based approaches, making it a go-to imputation method for a wide range of data types and dimensionalities. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 21:40:36 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Simkus",
"Vaidotas",
""
],
[
"Gutmann",
"Michael U.",
""
]
] |
2506.09259 | Rafal Kocielnik | Zhuofang Li, Rafal Kocielnik, Fereshteh Soltani, Penphob (Andrea) Boonyarungsrit, Animashree Anandkumar, R. Michael Alvarez | Self-Anchored Attention Model for Sample-Efficient Classification of Prosocial Text Chat | null | null | null | null | cs.CL cs.AI cs.CY | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Millions of players engage daily in competitive online games, communicating through in-game chat. Prior research has focused on detecting relatively small volumes of toxic content using various Natural Language Processing (NLP) techniques for the purpose of moderation. However, recent studies emphasize the importance of detecting prosocial communication, which can be as crucial as identifying toxic interactions. Recognizing prosocial behavior allows for its analysis, rewarding, and promotion. Unlike toxicity, there are limited datasets, models, and resources for identifying prosocial behaviors in game-chat text. In this work, we employed unsupervised discovery combined with game domain expert collaboration to identify and categorize prosocial player behaviors from game chat. We further propose a novel Self-Anchored Attention Model (SAAM) which gives 7.9% improvement compared to the best existing technique. The approach utilizes the entire training set as "anchors" to help improve model performance under the scarcity of training data. This approach led to the development of the first automated system for classifying prosocial behaviors in in-game chats, particularly given the low-resource settings where large-scale labeled data is not available. Our methodology was applied to one of the most popular online gaming titles - Call of Duty(R): Modern Warfare(R)II, showcasing its effectiveness. This research is novel in applying NLP techniques to discover and classify prosocial behaviors in player in-game chat communication. It can help shift the focus of moderation from solely penalizing toxicity to actively encouraging positive interactions on online platforms. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 21:40:54 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Li",
"Zhuofang",
"",
"Andrea"
],
[
"Kocielnik",
"Rafal",
"",
"Andrea"
],
[
"Soltani",
"Fereshteh",
"",
"Andrea"
],
[
"Penphob",
"",
"",
"Andrea"
],
[
"Boonyarungsrit",
"",
""
],
[
"Anandkumar",
"Animashree",
""
],
[
"Alvarez",
"R. Michael",
""
]
] |
2506.09260 | Yibin Lei | Yibin Lei, Tao Shen, Andrew Yates | ThinkQE: Query Expansion via an Evolving Thinking Process | null | null | null | null | cs.IR cs.CL | http://creativecommons.org/licenses/by/4.0/ | Effective query expansion for web search benefits from promoting both exploration and result diversity to capture multiple interpretations and facets of a query. While recent LLM-based methods have improved retrieval performance and demonstrate strong domain generalization without additional training, they often generate narrowly focused expansions that overlook these desiderata. We propose ThinkQE, a test-time query expansion framework addressing this limitation through two key components: a thinking-based expansion process that encourages deeper and comprehensive semantic exploration, and a corpus-interaction strategy that iteratively refines expansions using retrieval feedback from the corpus. Experiments on diverse web search benchmarks (DL19, DL20, and BRIGHT) show ThinkQE consistently outperforms prior approaches, including training-intensive dense retrievers and rerankers. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 21:41:01 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Lei",
"Yibin",
""
],
[
"Shen",
"Tao",
""
],
[
"Yates",
"Andrew",
""
]
] |
2506.09261 | Alessandro Della Corte | Alessandro Della Corte and Marco Farotti | Chains without regularity | null | null | null | null | math.DS | http://creativecommons.org/licenses/by/4.0/ | We study chain-recurrence and chain-transitivity in compact dynamical systems without any regularity assumptions. We prove that every compact system has a chain-recurrent point and a closed, invariant, chain-transitive subsystem. We also prove that, in fact, the chain-recurrent point whose existence is established is recurrent with respect to a stronger relation than chain-recurrence, which we identify. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 21:41:04 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Della Corte",
"Alessandro",
""
],
[
"Farotti",
"Marco",
""
]
] |
2506.09263 | David Jewitt | David Jewitt. Jing Li, Michael Jaeger and Yoonyoung Kim | Down But Not Out: The Case of Long-Period Comet C/2021 O3 (Panstarrs) | The Astronomical Journal, 26 pages, 8 figures | null | null | null | astro-ph.EP | http://creativecommons.org/licenses/by/4.0/ | We combine ground- and space-based observations of long-period comet C/2021 O3 (Panstarrs) (perihelion distance 0.287 au) in order to investigate its reported near-perihelion destruction. Pre-perihelion photometric observations show a remarkably small heliocentric dependence of the scattered light, $\propto r_H^{-s}$ with $s = 2.59\pm0.21$, distinct from values reported in other long-period comets, for which $s$ = 4 is the canonical standard. The index is smaller than expected of coma production by equilibrium sublimation of either supervolatiles (for which $s \sim$ 4 is expected), or water ice ($s \sim$ 6 to 8) across the $\sim$4 au to 2 au range. The absolute magnitude deduced from the pre-perihelion data is $H$ = 13.0$\pm$0.3 (coma scattering cross-section $\sim$225 km$^2$ for an assumed geometric albedo 0.04) while, after perihelion, the cross-section fades by a factor of 25 to $H$ = 16.5 ($\sim$9 km$^2$). STEREO spacecraft observations near perihelion show a long debris trail whose properties are consistent with forward scattering from radius $\sim$7 $\mu$m particles. The data show that the nucleus of C/2021 O3 was not destroyed at perihelion. Although the lightcurve from 3.9 au inbound to 0.8 au outbound cannot be uniquely interpreted, a simple and plausible explanation is provided by seasonal dimming on a nucleus having high obliquity and an asymmetric distribution of near-surface volatiles. The survival of the nucleus against rotational disruption suggests a pre-perihelion nucleus radius $r_n \gtrsim$ 1.0 km while the photometric limit to the radius of the nucleus after perihelion is $r_n < 1.7$ km (geometric albedo 0.04 assumed). | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 21:55:49 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Li",
"David Jewitt. Jing",
""
],
[
"Jaeger",
"Michael",
""
],
[
"Kim",
"Yoonyoung",
""
]
] |
2506.09264 | Robin Denart | Robin Denart | The bunkbed conjecture still holds for cactus graphs and for graphs with certain biconnected components | 15 pages | null | null | null | math.PR math.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recently, the bunkbed conjecture has been shown to be false, which naturally prompts questions on how to classify the graphs that still satisfy the conjecture. We distinguish between a weak version of the bunkbed conjecture where all the horizontal edges of the bunkbed graph are present with the same probability and a strong version of the conjecture where the edge weights on the underlying graph may be assigned individually. We show that any given graph satisfies either version of the conjecture if and only if all of its biconnected components do. Moreover, we show that all cactus graphs satisfy the strong version, and by combining previous results of other authors, any graph $G$ such that every biconnected component of $G$ is either a cycle, complete, complete bipartite, symmetric complete $k$-partite or an edge difference of a complete graph and a complete subgraph satisfies the weak version. Furthermore, we apply the aforementioned results to show that any counterexample to the strong version of the bunkbed conjecture contains a non-trivial subdivision of the diamond graph as a minor and demonstrate how this result might be strengthened in the future. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 21:56:51 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Denart",
"Robin",
""
]
] |
2506.09265 | Thuc Mai | Thuc T. Mai, Amber McCreary, K.F. Garrity, Rebecca L. Dally, Sambridhi Shah, Bryan C. Chakoumakos, Md Nasim Afroj Taj, Jeffrey W. Lynn, Michael A. McGuire, Benjamin S. Conner, Mona Zebarjadi, Janice L. Musfeldt, Angela R. Hight Walker, Rahul Rao, Michael A. Susner | Spin-lattice entanglement in $\mathbf{CoPS}_3$ | 12 pages, 5 figures | null | null | null | cond-mat.str-el cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | Complex chalcogenides in the $M$PS$_3$ family of materials ($M$ = Mn, Fe, Co, and Ni) display remarkably different phase progressions depending upon the metal center orbital filling, character of the P-P linkage, and size of the van der Waals gap. There is also a stacking pattern and spin state difference between the lighter and heavier transition metal-containing systems that places CoPS$_3$ at the nexus of these activities. Despite these unique properties, this compound is under-explored. Here, we bring together Raman scattering spectroscopy and infrared absorption spectroscopy with X-ray techniques to identify a structural component to the 119 K magnetic ordering transition as well as a remarkable lower temperature set of magnon-phonon pairs that engage in avoided crossings along with a magnetic scattering continuum that correlates with phonon lifetime effects. These findings point to strong spin-phonon entanglement as well as opportunities to control these effects under external stimuli. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 21:59:21 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Mai",
"Thuc T.",
""
],
[
"McCreary",
"Amber",
""
],
[
"Garrity",
"K. F.",
""
],
[
"Dally",
"Rebecca L.",
""
],
[
"Shah",
"Sambridhi",
""
],
[
"Chakoumakos",
"Bryan C.",
""
],
[
"Taj",
"Md Nasim Afroj",
""
],
[
"Lynn",
"Jeffrey W.",
""
],
[
"McGuire",
"Michael A.",
""
],
[
"Conner",
"Benjamin S.",
""
],
[
"Zebarjadi",
"Mona",
""
],
[
"Musfeldt",
"Janice L.",
""
],
[
"Walker",
"Angela R. Hight",
""
],
[
"Rao",
"Rahul",
""
],
[
"Susner",
"Michael A.",
""
]
] |
2506.09267 | Abhirup Datta | Abhirup Datta and Michael L. Stein | Consistent Infill Estimability of the Regression Slope Between Gaussian Random Fields Under Spatial Confounding | null | null | null | null | math.ST stat.TH | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The problem of estimating the slope parameter in regression between two spatial processes under confounding by an unmeasured spatial process has received widespread attention in the recent statistical literature. Yet, a fundamental question remains unsolved: when is this slope consistently estimable under spatial confounding, with existing insights being largely empirical or estimator-specific. In this manuscript, we characterize conditions for consistent estimability of the regression slope between Gaussian random fields (GRFs). Under fixed-domain (infill) asymptotics, we give sufficient conditions for consistent estimability using a novel characterization of the regression slope as the ratio of principal irregular terms of covariances, dictating the relative local behavior of the exposure and confounder processes. When estimability holds, we provide consistent estimators of the slope using local differencing (taking discrete differences or Laplacians of the processes of suitable order). Using functional analysis results on Paley-Wiener spaces, we then provide an easy-to-verify necessary condition for consistent estimability of the slope in terms of the relative spectral tail decays of the confounder and exposure. As a by-product, we establish a novel and general spectral condition on the equivalence of measures on the paths of multivariate GRFs with component fields of varying smoothnesses, a result of independent importance. We show that for the Mat\'ern, power-exponential, generalized Cauchy, and coregionalization families, the necessary and sufficient conditions become identical, thereby providing a complete characterization of consistent estimability of the slope under spatial confounding. The results are extended to accommodate measurement error using local-averaging-and-differencing based estimators. Finite sample behavior is explored via numerical experiments. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 22:04:30 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Datta",
"Abhirup",
""
],
[
"Stein",
"Michael L.",
""
]
] |
2506.09268 | Henri Alam | Henri Alam and Antonio de Domenico and Tareq Si Salem and Florian Kaltenberger | A Multi-Armed Bandit Framework for Online Optimisation in Green Integrated Terrestrial and Non-Terrestrial Networks | To be published in 2025 IEEE International Workshop on Signal Processing and Artificial Intelligence in Wireless Communications (IEEE SPAWC 2025) | null | null | null | cs.NI cs.AI | http://creativecommons.org/licenses/by/4.0/ | Integrated terrestrial and non-terrestrial network (TN-NTN) architectures offer a promising solution for expanding coverage and improving capacity for the network. While non-terrestrial networks (NTNs) are primarily exploited for these specific reasons, their role in alleviating terrestrial network (TN) load and enabling energy-efficient operation has received comparatively less attention. In light of growing concerns associated with the densification of terrestrial deployments, this work aims to explore the potential of NTNs in supporting a more sustainable network. In this paper, we propose a novel online optimisation framework for integrated TN-NTN architectures, built on a multi-armed bandit (MAB) formulation and leveraging the Bandit-feedback Constrained Online Mirror Descent (BCOMD) algorithm. Our approach adaptively optimises key system parameters--including bandwidth allocation, user equipment (UE) association, and macro base station (MBS) shutdown--to balance network capacity and energy efficiency in real time. Extensive system-level simulations over a 24-hour period show that our framework significantly reduces the proportion of unsatisfied UEs during peak hours and achieves up to 19% throughput gains and 5% energy savings in low-traffic periods, outperforming standard network settings following 3GPP recommendations. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 22:06:15 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Alam",
"Henri",
""
],
[
"de Domenico",
"Antonio",
""
],
[
"Salem",
"Tareq Si",
""
],
[
"Kaltenberger",
"Florian",
""
]
] |
2506.09269 | Hong Duc Bui | Hong Duc Bui | Straight-line Orthogonal Drawing of Complete Ternary Tree Requires $O(n^{1.032})$ Area | 7 pages, 4 figures | null | null | null | cs.CG | http://creativecommons.org/licenses/by/4.0/ | We resolve a conjecture posed by Covella, Frati and Patrignani by proving the straight-line orthogonal drawing of the complete ternary tree with $n$ nodes satisfying the subtree separation property with smallest area has area $\Omega(n^{1.031})$. We also improve the upper bound of this area to $O(n^{1.032})$. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 22:06:57 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Bui",
"Hong Duc",
""
]
] |
2506.09270 | Rodrigo Carrasco-Davis | Rodrigo Carrasco-Davis, Sebastian Lee, Claudia Clopath, Will Dabney | Uncertainty Prioritized Experience Replay | Accepted at Reinforcement Learning Conference | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Prioritized experience replay, which improves sample efficiency by selecting relevant transitions to update parameter estimates, is a crucial component of contemporary value-based deep reinforcement learning models. Typically, transitions are prioritized based on their temporal difference error. However, this approach is prone to favoring noisy transitions, even when the value estimation closely approximates the target mean. This phenomenon resembles the noisy TV problem postulated in the exploration literature, in which exploration-guided agents get stuck by mistaking noise for novelty. To mitigate the disruptive effects of noise in value estimation, we propose using epistemic uncertainty estimation to guide the prioritization of transitions from the replay buffer. Epistemic uncertainty quantifies the uncertainty that can be reduced by learning, hence reducing transitions sampled from the buffer generated by unpredictable random processes. We first illustrate the benefits of epistemic uncertainty prioritized replay in two tabular toy models: a simple multi-arm bandit task, and a noisy gridworld. Subsequently, we evaluate our prioritization scheme on the Atari suite, outperforming quantile regression deep Q-learning benchmarks; thus forging a path for the use of uncertainty prioritized replay in reinforcement learning agents. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 22:07:13 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Carrasco-Davis",
"Rodrigo",
""
],
[
"Lee",
"Sebastian",
""
],
[
"Clopath",
"Claudia",
""
],
[
"Dabney",
"Will",
""
]
] |
2506.09271 | Oskar Gustafsson | Oskar Gustafsson and Mattias Villani | Variational inference for steady-state BVARs | Working paper. Will be updated with additional results and submitted to a journal | null | null | null | stat.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The steady-state Bayesian vector autoregression (BVAR) makes it possible to incorporate prior information about the long-run mean of the process. This has been shown in many studies to substantially improve forecasting performance, and the model is routinely used for forecasting and macroeconomic policy analysis at central banks and other financial institutions. Steady-steady BVARs are estimated using Gibbs sampling, which is time-consuming for the increasingly popular large-scale BVAR models with many variables. We propose a fast variational inference (VI) algorithm for approximating the parameter posterior and predictive distribution of the steady-state BVAR, as well as log predictive scores for model comparison. We use simulated and real US macroeconomic data to show that VI produces results that are very close to those from Gibbs sampling. The computing time of VI can be orders of magnitude lower than Gibbs sampling, in particular for log predictive scores, and VI is shown to scale much better with the number of time series in the system. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 22:09:00 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Gustafsson",
"Oskar",
""
],
[
"Villani",
"Mattias",
""
]
] |
2506.09272 | Samuel Holt | Samuel Holt, Max Ruiz Luyten, Antonin Berthon, Mihaela van der Schaar | G-Sim: Generative Simulations with Large Language Models and Gradient-Free Calibration | Accepted at the 42nd International Conference on Machine Learning (ICML 2025). 9 pages, 3 figures | Proceedings of the 42nd International Conference on Machine Learning, Vancouver, Canada. PMLR 267, 2025 | null | null | cs.LG stat.ML | http://creativecommons.org/licenses/by/4.0/ | Constructing robust simulators is essential for asking "what if?" questions and guiding policy in critical domains like healthcare and logistics. However, existing methods often struggle, either failing to generalize beyond historical data or, when using Large Language Models (LLMs), suffering from inaccuracies and poor empirical alignment. We introduce G-Sim, a hybrid framework that automates simulator construction by synergizing LLM-driven structural design with rigorous empirical calibration. G-Sim employs an LLM in an iterative loop to propose and refine a simulator's core components and causal relationships, guided by domain knowledge. This structure is then grounded in reality by estimating its parameters using flexible calibration techniques. Specifically, G-Sim can leverage methods that are both likelihood-free and gradient-free with respect to the simulator, such as gradient-free optimization for direct parameter estimation or simulation-based inference for obtaining a posterior distribution over parameters. This allows it to handle non-differentiable and stochastic simulators. By integrating domain priors with empirical evidence, G-Sim produces reliable, causally-informed simulators, mitigating data-inefficiency and enabling robust system-level interventions for complex decision-making. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 22:14:34 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Holt",
"Samuel",
""
],
[
"Luyten",
"Max Ruiz",
""
],
[
"Berthon",
"Antonin",
""
],
[
"van der Schaar",
"Mihaela",
""
]
] |
2506.09273 | Shimin Wang | Telema Harry and Martin Guay and Shimin Wang and Richard D. Braatz | Data-Driven Nonlinear Regulation: Gaussian Process Learning | null | null | null | null | eess.SY cs.DM cs.SY math.OC nlin.AO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This article addresses the output regulation problem for a class of nonlinear systems using a data-driven approach. An output feedback controller is proposed that integrates a traditional control component with a data-driven learning algorithm based on Gaussian Process (GP) regression to learn the nonlinear internal model. Specifically, a data-driven technique is employed to directly approximate the unknown internal model steady-state map from observed input-output data online. Our method does not rely on model-based observers utilized in previous studies, making it robust and suitable for systems with modelling errors and model uncertainties. Finally, we demonstrate through numerical examples and detailed stability analysis that, under suitable conditions, the closed-loop system remains bounded and converges to a compact set, with the size of this set decreasing as the accuracy of the data-driven model improves over time. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 22:21:28 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Harry",
"Telema",
""
],
[
"Guay",
"Martin",
""
],
[
"Wang",
"Shimin",
""
],
[
"Braatz",
"Richard D.",
""
]
] |
2506.09274 | Pietro Conte | P. Conte, G. Consolati, M. Prata, M. Berghold, R. Caravita, R. Ferguson, M. Grosbart, F. Guatieri, S. Haider, G. Khatri, L. Lappo, P. Moskal, M. M\"unster, L. Penasa and S. Sharma | AEgIS Experiment at CERN: Design and Commissioning of SARA (Scintillator Assemblies to Reveal Annihilations) | 16 pages, 12 figures | null | null | null | physics.ins-det hep-ex | http://creativecommons.org/licenses/by/4.0/ | SARA is the system of plastic scintillators coupled with silicon photomultipliers that will take part in the AEgIS experiment at CERN, measuring the time-of-flight of antihydrogen as it falls through a moir\'e deflectometer. Its development focused on simplicity, versatility and economy of the design and was supported by both physical tests and finite elements numerical simulations. The instrument's structure pairs the utilization of the scintillators as structural components with custom made 3D printed corner elements and the electronics allows to select between coincidence discrimination made on each scintillator and made between different scintillators. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 22:24:30 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Conte",
"P.",
""
],
[
"Consolati",
"G.",
""
],
[
"Prata",
"M.",
""
],
[
"Berghold",
"M.",
""
],
[
"Caravita",
"R.",
""
],
[
"Ferguson",
"R.",
""
],
[
"Grosbart",
"M.",
""
],
[
"Guatieri",
"F.",
""
],
[
"Haider",
"S.",
""
],
[
"Khatri",
"G.",
""
],
[
"Lappo",
"L.",
""
],
[
"Moskal",
"P.",
""
],
[
"Münster",
"M.",
""
],
[
"Penasa",
"L.",
""
],
[
"Sharma",
"S.",
""
]
] |
2506.09275 | Jonas Svedas | Jonas Svedas, Hannah Watson, Nathan Laubeuf, Diksha Moolchandani, Abubakr Nada, Arjun Singh, Dwaipayan Biswas, James Myers, Debjyoti Bhattacharjee | A Survey of End-to-End Modeling for Distributed DNN Training: Workloads, Simulators, and TCO | null | null | null | null | cs.DC | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Distributed deep neural networks (DNNs) have become a cornerstone for scaling machine learning to meet the demands of increasingly complex applications. However, the rapid growth in model complexity far outpaces CMOS technology scaling, making sustainable and efficient system design a critical challenge. Addressing this requires coordinated co-design across software, hardware, and technology layers. Due to the prohibitive cost and complexity of deploying full-scale training systems, simulators play a pivotal role in enabling this design exploration. This survey reviews the landscape of distributed DNN training simulators, focusing on three major dimensions: workload representation, simulation infrastructure, and models for total cost of ownership (TCO) including carbon emissions. It covers how workloads are abstracted and used in simulation, outlines common workload representation methods, and includes comprehensive comparison tables covering both simulation frameworks and TCO/emissions models, detailing their capabilities, assumptions, and areas of focus. In addition to synthesizing existing tools, the survey highlights emerging trends, common limitations, and open research challenges across the stack. By providing a structured overview, this work supports informed decision-making in the design and evaluation of distributed training systems. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 22:25:29 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Svedas",
"Jonas",
""
],
[
"Watson",
"Hannah",
""
],
[
"Laubeuf",
"Nathan",
""
],
[
"Moolchandani",
"Diksha",
""
],
[
"Nada",
"Abubakr",
""
],
[
"Singh",
"Arjun",
""
],
[
"Biswas",
"Dwaipayan",
""
],
[
"Myers",
"James",
""
],
[
"Bhattacharjee",
"Debjyoti",
""
]
] |
2506.09276 | Lorenzo Steccanella | Lorenzo Steccanella, Joshua B. Evans, \"Ozg\"ur \c{S}im\c{s}ek, Anders Jonsson | Learning The Minimum Action Distance | null | null | null | null | cs.LG cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This paper presents a state representation framework for Markov decision processes (MDPs) that can be learned solely from state trajectories, requiring neither reward signals nor the actions executed by the agent. We propose learning the minimum action distance (MAD), defined as the minimum number of actions required to transition between states, as a fundamental metric that captures the underlying structure of an environment. MAD naturally enables critical downstream tasks such as goal-conditioned reinforcement learning and reward shaping by providing a dense, geometrically meaningful measure of progress. Our self-supervised learning approach constructs an embedding space where the distances between embedded state pairs correspond to their MAD, accommodating both symmetric and asymmetric approximations. We evaluate the framework on a comprehensive suite of environments with known MAD values, encompassing both deterministic and stochastic dynamics, as well as discrete and continuous state spaces, and environments with noisy observations. Empirical results demonstrate that the proposed approach not only efficiently learns accurate MAD representations across these diverse settings but also significantly outperforms existing state representation methods in terms of representation quality. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 22:27:11 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Steccanella",
"Lorenzo",
""
],
[
"Evans",
"Joshua B.",
""
],
[
"Şimşek",
"Özgür",
""
],
[
"Jonsson",
"Anders",
""
]
] |
2506.09278 | Nikhil Keetha | Yuchen Zhang, Nikhil Keetha, Chenwei Lyu, Bhuvan Jhamb, Yutian Chen, Yuheng Qiu, Jay Karhade, Shreyas Jha, Yaoyu Hu, Deva Ramanan, Sebastian Scherer, Wenshan Wang | UFM: A Simple Path towards Unified Dense Correspondence with Flow | Project Page: https://uniflowmatch.github.io/ | null | null | null | cs.CV cs.LG cs.RO | http://creativecommons.org/licenses/by/4.0/ | Dense image correspondence is central to many applications, such as visual odometry, 3D reconstruction, object association, and re-identification. Historically, dense correspondence has been tackled separately for wide-baseline scenarios and optical flow estimation, despite the common goal of matching content between two images. In this paper, we develop a Unified Flow & Matching model (UFM), which is trained on unified data for pixels that are co-visible in both source and target images. UFM uses a simple, generic transformer architecture that directly regresses the (u,v) flow. It is easier to train and more accurate for large flows compared to the typical coarse-to-fine cost volumes in prior work. UFM is 28% more accurate than state-of-the-art flow methods (Unimatch), while also having 62% less error and 6.7x faster than dense wide-baseline matchers (RoMa). UFM is the first to demonstrate that unified training can outperform specialized approaches across both domains. This result enables fast, general-purpose correspondence and opens new directions for multi-modal, long-range, and real-time correspondence tasks. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 22:32:13 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Zhang",
"Yuchen",
""
],
[
"Keetha",
"Nikhil",
""
],
[
"Lyu",
"Chenwei",
""
],
[
"Jhamb",
"Bhuvan",
""
],
[
"Chen",
"Yutian",
""
],
[
"Qiu",
"Yuheng",
""
],
[
"Karhade",
"Jay",
""
],
[
"Jha",
"Shreyas",
""
],
[
"Hu",
"Yaoyu",
""
],
[
"Ramanan",
"Deva",
""
],
[
"Scherer",
"Sebastian",
""
],
[
"Wang",
"Wenshan",
""
]
] |
2506.09279 | Ziyi Chen | Ziyi Chen, Yiyang Liu, Mattia Prosperi, Krishna Vaddiparti, Robert L Cook, Jiang Bian, Yi Guo, Yonghui Wu | A Topic Modeling Analysis of Stigma Dimensions, Social, and Related Behavioral Circumstances in Clinical Notes Among Patients with HIV | null | null | null | null | cs.LG cs.IR | http://creativecommons.org/licenses/by-sa/4.0/ | Objective: To characterize stigma dimensions, social, and related behavioral circumstances in people living with HIV (PLWHs) seeking care, using natural language processing methods applied to a large collection of electronic health record (EHR) clinical notes from a large integrated health system in the southeast United States. Methods: We identified 9,140 cohort of PLWHs from the UF Health IDR and performed topic modeling analysis using Latent Dirichlet Allocation (LDA) to uncover stigma dimensions, social, and related behavioral circumstances. Domain experts created a seed list of HIV-related stigma keywords, then applied a snowball strategy to iteratively review notes for additional terms until saturation was reached. To identify more target topics, we tested three keyword-based filtering strategies. Domain experts manually reviewed the detected topics using the prevalent terms and key discussion topics. Word frequency analysis was used to highlight the prevalent terms associated with each topic. In addition, we conducted topic variation analysis among subgroups to examine differences across age and sex-specific demographics. Results and Conclusion: Topic modeling on sentences containing at least one keyword uncovered a wide range of topic themes associated with HIV-related stigma, social, and related behaviors circumstances, including "Mental Health Concern and Stigma", "Social Support and Engagement", "Limited Healthcare Access and Severe Illness", "Treatment Refusal and Isolation" and so on. Topic variation analysis across age subgroups revealed differences. Extracting and understanding the HIV-related stigma dimensions, social, and related behavioral circumstances from EHR clinical notes enables scalable, time-efficient assessment, overcoming the limitations of traditional questionnaires and improving patient outcomes. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 22:35:49 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Chen",
"Ziyi",
""
],
[
"Liu",
"Yiyang",
""
],
[
"Prosperi",
"Mattia",
""
],
[
"Vaddiparti",
"Krishna",
""
],
[
"Cook",
"Robert L",
""
],
[
"Bian",
"Jiang",
""
],
[
"Guo",
"Yi",
""
],
[
"Wu",
"Yonghui",
""
]
] |
2506.09280 | Haitian Jiang | Haitian Jiang, Shaowei Zhu, Zhen Zhang, Zhenyu Song, Xinwei Fu, Zhen Jia, Yida Wang, Jinyang Li | TTrace: Lightweight Error Checking and Diagnosis for Distributed Training | null | null | null | null | cs.DC cs.LG cs.NA math.NA | http://creativecommons.org/licenses/by/4.0/ | Distributed training is essential for scaling the training of large neural network models, such as large language models (LLMs), across thousands of GPUs. However, the complexity of distributed training programs makes them particularly prone to silent bugs, which do not produce explicit error signal but lead to incorrect training outcome. Effectively detecting and localizing such silent bugs in distributed training is challenging. Common debugging practice using metrics like training loss or gradient norm curves can be inefficient and ineffective. Additionally, obtaining intermediate tensor values and determining whether they are correct during silent bug localization is difficult, particularly in the context of low-precision training.
To address those challenges, we design and implement TTrace, the first system capable of detecting and localizing silent bugs in distributed training. TTrace collects intermediate tensors from distributing training in a fine-grained manner and compares them against those from a trusted single-device reference implementation. To properly compare the floating-point values in the tensors, we propose novel mathematical analysis that provides a guideline for setting thresholds, enabling TTrace to distinguish bug-induced errors from floating-point round-off errors. Experimental results demonstrate that TTrace effectively detects 11 existing bugs and 3 new bugs in the widely used Megatron-LM framework, while requiring fewer than 10 lines of code change. TTrace is effective in various training recipes, including low-precision recipes involving BF16 and FP8. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 22:39:14 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Jiang",
"Haitian",
""
],
[
"Zhu",
"Shaowei",
""
],
[
"Zhang",
"Zhen",
""
],
[
"Song",
"Zhenyu",
""
],
[
"Fu",
"Xinwei",
""
],
[
"Jia",
"Zhen",
""
],
[
"Wang",
"Yida",
""
],
[
"Li",
"Jinyang",
""
]
] |
2506.09281 | Juan Maga\~na | Dorian Araya, Cristian Castillo, Genly Leon, Juan Maga\~na, Angie Barr Dom\'inguez and Miguel A. Garc\'ia-Aspeitia | $w_{dm}$-$w_{de}$ cosmological model with new data samples of cosmological observations | 22 pages, 9 figures, 12 tables, submitted to PDU | null | null | null | astro-ph.CO gr-qc | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We revisit a cosmological model where dark matter (DM) and dark energy (DE) follow barotropic equations of state, allowing deviations from the standard $\Lambda$CDM framework (i.e. $w_{dm} \neq 0$, $w_{de} \neq -1$) considering both, flat and non-flat curvature. Using a dynamical system approach, we identify equilibrium states that govern stability, expansion, and contraction. Expansion occurs when $H>0$, while contraction is linked to $H < 0$. Accelerated expansion arises from DE dominance, whereas radiation- and matter-dominated phases lead to deceleration. Some solutions are unphysical, because of density constraints, but viable cases offer insights into cosmic transitions, including the Einstein static universe, which allows shifts between accelerating and decelerating phases. We perform a Bayesian analysis with updated datasets, including observational Hubble data, Pantheon+ Type Ia supernovae, strong lensing systems, and baryon acoustic oscillations, to constrain the parameters $w_{dm}$ and $w_{de}$. Our results from the data joint analysis show consistency with $\Lambda$CDM within $3\sigma$, but none of the cases reproduce $w_{dm} = 0$ and $w_{de} = -1$. Nevertheless, the comparison with the standard model using the Akaike and Bayesian information criteria indicates that only the non-flat scenario has the potential to be competitive. This suggests that a non-dust-like DM may impact structure formation, while DE could shift toward quintessence fluid. While $\Lambda$CDM remains a strong model, our findings indicate that alternative dark sector models with non-standard EoS could be viable and offer new insights into cosmic evolution. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 22:45:36 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Araya",
"Dorian",
""
],
[
"Castillo",
"Cristian",
""
],
[
"Leon",
"Genly",
""
],
[
"Magaña",
"Juan",
""
],
[
"Domínguez",
"Angie Barr",
""
],
[
"García-Aspeitia",
"Miguel A.",
""
]
] |
2506.09282 | Dhruv Parikh | Dhruv Parikh, Viktor Prasanna | ScalableHD: Scalable and High-Throughput Hyperdimensional Computing Inference on Multi-Core CPUs | IC3 | null | null | null | cs.DC cs.LG | http://creativecommons.org/licenses/by/4.0/ | Hyperdimensional Computing (HDC) is a brain-inspired computing paradigm that represents and manipulates information using high-dimensional vectors, called hypervectors (HV). Traditional HDC methods, while robust to noise and inherently parallel, rely on single-pass, non-parametric training and often suffer from low accuracy. To address this, recent approaches adopt iterative training of base and class HVs, typically accelerated on GPUs. Inference, however, remains lightweight and well-suited for real-time execution. Yet, efficient HDC inference has been studied almost exclusively on specialized hardware such as FPGAs and GPUs, with limited attention to general-purpose multi-core CPUs. To address this gap, we propose ScalableHD for scalable and high-throughput HDC inference on multi-core CPUs. ScalableHD employs a two-stage pipelined execution model, where each stage is parallelized across cores and processes chunks of base and class HVs. Intermediate results are streamed between stages using a producer-consumer mechanism, enabling on-the-fly consumption and improving cache locality. To maximize performance, ScalableHD integrates memory tiling and NUMA-aware worker-to-core binding. Further, it features two execution variants tailored for small and large batch sizes, each designed to exploit compute parallelism based on workload characteristics while mitigating the memory-bound compute pattern that limits HDC inference performance on modern multi-core CPUs. ScalableHD achieves up to 10x speedup in throughput (samples per second) over state-of-the-art baselines such as TorchHD, across a diverse set of tasks ranging from human activity recognition to image classification, while preserving task accuracy. Furthermore, ScalableHD exhibits robust scalability: increasing the number of cores yields near-proportional throughput improvements. | [
{
"version": "v1",
"created": "Tue, 10 Jun 2025 22:46:12 GMT"
}
] | 2025-06-12T00:00:00 | [
[
"Parikh",
"Dhruv",
""
],
[
"Prasanna",
"Viktor",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.