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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
2506.14348 | Olena Kompaniiets | O. V. Kompaniiets, A. A. Vasylenko, I. B. Vavilova | The 2MIG isolated AGNs -- 2. X-ray general properties and peculiarities | 11 pages, 13 figures | null | null | null | astro-ph.HE astro-ph.GA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We have analyzed a sample of 2MIG isolated galaxies hosting AGNs (isolated AGNs) to assess whether their nuclear activity differs from that in a denser environment. The isolation criteria rule out interactions with other galaxies of similar size for at least $\sim$3 Gyr. We systemized the available Swift, NuSTAR, XMM-Newton, Chandra, and INTEGRAL X-ray data for isolated AGN at $z<0.05$, determining their general X-ray properties and peculiarities, spectral models, SMBH masses, and possible parameter correlations. We investigated the best spectral models of 25 isolated AGNs, including 10 recently fitted in this work. The intrinsic luminosities $\log L_{2-10,\mathrm{keV}}$ do not exceed $\sim43$.
Our results indicate that the isolation of galaxies in the nearby Universe does not significantly affect nuclear activity. This is confirmed by the diversity of accretion types and the absence of any advantage of one particular basic/compound spectral model. We note an interesting case discovered for the ESO499-041 spectrum, where the data-to-model ratio shows significant changes after 6.4 keV. These deviations from the continuum may indicate the presence of a relativistic Iron line in the Chandra spectrum of ESO 499-041 compared to the Swift spectrum. Most isolated AGNs contain SMBH with $M_{\mathrm{SMBH}} \leq 10^7,M_{\odot}$. For the first time, we found a linear correlation between $L_{2-10,\mathrm{keV}}$ and $\log M_{\mathrm{SMBH}}$ that is not observed in other AGN samples. This trend may be affected by the limited sample size and requires further confirmation with larger datasets. Finally, we identified UGC10120, NG6300, and CGCG243-024 as promising candidates for Milky Way analogues. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 09:44:31 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Kompaniiets",
"O. V.",
""
],
[
"Vasylenko",
"A. A.",
""
],
[
"Vavilova",
"I. B.",
""
]
] |
2506.14349 | Mauritz Niklas Cartier Van Dissel | Mauritz N. Cartier van Dissel, Samuel Martin-Gutierrez, Lisette Esp\'in-Noboa, Ana Mar\'ia Jaramillo, Fariba Karimi | hyperFA*IR: A hypergeometric approach to fair rankings with finite candidate pool | In Proceedings of the 2025 ACM Conference on Fairness, Accountability, and Transparency (FAccT'25) | null | 10.1145/3715275.3732143 | null | cs.CY cs.IR stat.AP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Ranking algorithms play a pivotal role in decision-making processes across diverse domains, from search engines to job applications. When rankings directly impact individuals, ensuring fairness becomes essential, particularly for groups that are marginalised or misrepresented in the data. Most of the existing group fairness frameworks often rely on ensuring proportional representation of protected groups. However, these approaches face limitations in accounting for the stochastic nature of ranking processes or the finite size of candidate pools. To this end, we present hyperFA*IR, a framework for assessing and enforcing fairness in rankings drawn from a finite set of candidates. It relies on a generative process based on the hypergeometric distribution, which models real-world scenarios by sampling without replacement from fixed group sizes. This approach improves fairness assessment when top-$k$ selections are large relative to the pool or when protected groups are small. We compare our approach to the widely used binomial model, which treats each draw as independent with fixed probability, and demonstrate$-$both analytically and empirically$-$that our method more accurately reproduces the statistical properties of sampling from a finite population. To operationalise this framework, we propose a Monte Carlo-based algorithm that efficiently detects unfair rankings by avoiding computationally expensive parameter tuning. Finally, we adapt our generative approach to define affirmative action policies by introducing weights into the sampling process. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 09:45:08 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"van Dissel",
"Mauritz N. Cartier",
""
],
[
"Martin-Gutierrez",
"Samuel",
""
],
[
"Espín-Noboa",
"Lisette",
""
],
[
"Jaramillo",
"Ana María",
""
],
[
"Karimi",
"Fariba",
""
]
] |
2506.14350 | Zoubida Ameur | Zoubida Ameur, Fr\'ed\'eric Lefebvre, Philippe De Lagrange and Milo\v{s} Radosavljevi\'c | FGA-NN: Film Grain Analysis Neural Network | null | null | null | null | cs.CV eess.IV | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Film grain, once a by-product of analog film, is now present in most cinematographic content for aesthetic reasons. However, when such content is compressed at medium to low bitrates, film grain is lost due to its random nature. To preserve artistic intent while compressing efficiently, film grain is analyzed and modeled before encoding and synthesized after decoding. This paper introduces FGA-NN, the first learning-based film grain analysis method to estimate conventional film grain parameters compatible with conventional synthesis. Quantitative and qualitative results demonstrate FGA-NN's superior balance between analysis accuracy and synthesis complexity, along with its robustness and applicability. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 09:48:27 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Ameur",
"Zoubida",
""
],
[
"Lefebvre",
"Frédéric",
""
],
[
"De Lagrange",
"Philippe",
""
],
[
"Radosavljević",
"Miloš",
""
]
] |
2506.14351 | Satyajit Guin | Keshab Chandra Bakshi, Satyajit Guin, Guruprasad | A noncommutative construction of families of biunitary matrices and application to subfactors | null | null | null | null | math.OA math.FA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We introduce a construction that, given a pair (u,v) of complex Hadamard matrices of the same order, generates infinitely many biunitary matrices of varying (and distinct) orders. As a key application, this framework yields nested sequences of vertex model subfactors that are not a tower of downward basic construction. Notably, the construction is noncommutative: interchanging the matrices (i.e., considering (v,u) instead of (u,v)) can lead to non-isomorphic subfactors. Focusing on the Hadamard equivalence class of the Fourier matrix, we provide a full characterization of the resulting vertex model subfactors, along with explicit computations of their relative commutants. Along the way, we conduct a detailed study of certain naturally arising inner and outer automorphisms that play a key role in the structure of these subfactors. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 09:48:40 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Bakshi",
"Keshab Chandra",
""
],
[
"Guin",
"Satyajit",
""
],
[
"Guruprasad",
"",
""
]
] |
2506.14352 | Philipp H\"onicke | Marie-Louis Venzke, Katja Frenzel and Philipp H\"onicke | Experimental determination of the sodium K-shell atomic fundamental parameters for X-ray spectroscopy | null | null | null | null | physics.atom-ph physics.app-ph | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Reliable fundamental parameters are essential for accurate X-ray fluorescence analysis of sodium. This work presents updated experimental values with reliable uncertainties for sodium K-shell fundamental parameters such as: fluorescence yield, photoionization cross sections, and Auger yields. Using a physically calibrated setup and a thin NaCl layer on a silicon nitride membrane, a holistic determination approach was applied to reduce uncertainties. The new values represent a significant improvement on the ones widely used in databases. All data are available via Zenodo to support precise sodium quantification in scientific and industrial applications. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 09:49:11 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Venzke",
"Marie-Louis",
""
],
[
"Frenzel",
"Katja",
""
],
[
"Hönicke",
"Philipp",
""
]
] |
2506.14353 | C\'edric Simal | C\'edric Simal, Julien Petit and Timoteo Carletti | The Shortest-Path distance on graphons | 37 pages, 3 figures | null | null | null | math.CO | http://creativecommons.org/licenses/by/4.0/ | We define an analogue of the shortest-path distance for graphons. The proposed method is rooted on the extension to graphons of Varadhan's formula, a result that links the solution of the heat equation on a Riemannian manifold to its geodesic distance. The resulting metric is integer-valued, and for step graphons obtained from finite graphs it is essentially equivalent to the usual shortest-path distance. We further draw a link between the Varadhan distance and the communicability distance, that contains information from all paths, not just shortest-paths, and thus provides a finer distance on graphons along with a natural isometric embedding into a Hilbert space. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 09:49:52 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Simal",
"Cédric",
""
],
[
"Petit",
"Julien",
""
],
[
"Carletti",
"Timoteo",
""
]
] |
2506.14354 | Taiki Ikeda | Taiki Ikeda, Sugumi Kanno, Jiro Soda | Enhancing photon-axion conversion probability with squeezed coherent states | 16 pages | null | null | KOBE-COSMO-25-11 | quant-ph gr-qc hep-ph hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In particle physics, axions and axion-like particles are ubiquitous. Remarkably, ultra-light axions could constitute dark matter or dark energy. Therefore, it is important to detect axions experimentally. In the presence of a magnetic field, a photon can be converted into an axion, and vice versa. Utilizing the conversion phenomenon, several methods for detecting axions have been proposed. To improve detectability, it is desirable to use quantum sensing. However, since the conversion process is usually treated as classical wave dynamics, it is unclear how to incorporate quantum effects such as entanglement. In this work, we formulate the photon-axion conversion in a quantum field theoretical manner. As a result, we succeed in evaluating the conversion probability from a photon quantum state to an axion quantum state. In particular, it turns out that squeezed coherent states with N photons added can significantly enhance the conversion probability. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 09:50:10 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Ikeda",
"Taiki",
""
],
[
"Kanno",
"Sugumi",
""
],
[
"Soda",
"Jiro",
""
]
] |
2506.14355 | Radmehr Jafari | Saeid Ansari, R. Jafari, Alireza Akbari, and Mehdi Abdi | Scaling and Universality at Noisy Quench Dynamical Quantum Phase Transitions | null | null | null | null | cond-mat.stat-mech quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Dynamical quantum phase transitions (DQPTs) have been studied in the extended XY model under both noiseless and noisy linear driven staggered field cases. In the time-independent staggered field case, the model exhibits a single critical point where the transition occurs from the spin-liquid phase to the antiferromagnetic phase. In the noiseless ramp case, unlike the transverse field XY model where DQPT always occurs for a quench crossing the single critical point, there is a critical sweep velocity above which the kinks corresponding to a DQPT are completely removed. Furthermore, in this case there are only two critical modes whose excitation probability is one-half. In the presence of a Gaussian white noise, we find that this critical sweep velocity decreases by increasing the noise strength, and scales linearly with the square of the noise intensity. A surprising result occurs when the noise intensity and sweep velocity are about the same order of magnitude, the number of critical modes is significantly increased, signalling a region with multiple critical modes. Furthermore, our findings indicate that the scaling of the dynamical free energy near the DQPTs time is the same for both noiseless and noisy ramp quenches. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 09:50:51 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Ansari",
"Saeid",
""
],
[
"Jafari",
"R.",
""
],
[
"Akbari",
"Alireza",
""
],
[
"Abdi",
"Mehdi",
""
]
] |
2506.14356 | Xiaoqi Wang | Xiaoqi Wang, Yi Wang, Lap-Pui Chau | EVA02-AT: Egocentric Video-Language Understanding with Spatial-Temporal Rotary Positional Embeddings and Symmetric Optimization | null | null | null | null | cs.CV cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Egocentric video-language understanding demands both high efficiency and accurate spatial-temporal modeling. Existing approaches face three key challenges: 1) Excessive pre-training cost arising from multi-stage pre-training pipelines, 2) Ineffective spatial-temporal encoding due to manually split 3D rotary positional embeddings that hinder feature interactions, and 3) Imprecise learning objectives in soft-label multi-instance retrieval, which neglect negative pair correlations. In this paper, we introduce EVA02-AT, a suite of EVA02-based video-language foundation models tailored to egocentric video understanding tasks. EVA02-AT first efficiently transfers an image-based CLIP model into a unified video encoder via a single-stage pretraining. Second, instead of applying rotary positional embeddings to isolated dimensions, we introduce spatial-temporal rotary positional embeddings along with joint attention, which can effectively encode both spatial and temporal information on the entire hidden dimension. This joint encoding of spatial-temporal features enables the model to learn cross-axis relationships, which are crucial for accurately modeling motion and interaction in videos. Third, focusing on multi-instance video-language retrieval tasks, we introduce the Symmetric Multi-Similarity (SMS) loss and a novel training framework that advances all soft labels for both positive and negative pairs, providing a more precise learning objective. Extensive experiments on Ego4D, EPIC-Kitchens-100, and Charades-Ego under zero-shot and fine-tuning settings demonstrate that EVA02-AT achieves state-of-the-art performance across diverse egocentric video-language tasks with fewer parameters. Models with our SMS loss also show significant performance gains on multi-instance retrieval benchmarks. Our code and models are publicly available at https://github.com/xqwang14/EVA02-AT . | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 09:51:51 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Wang",
"Xiaoqi",
""
],
[
"Wang",
"Yi",
""
],
[
"Chau",
"Lap-Pui",
""
]
] |
2506.14357 | Yury Arlinskii | Yury Arlinskii and Konrad Schm\"udgen | $C$-self-adjoint contractive extensions of $C$-symmetric non-densely defined contractions | 18 pages | null | null | null | math.FA | http://creativecommons.org/licenses/by/4.0/ | Given a conjugation (involution) $C$ on a Hilbert space, $C$-self-adjoint contractive extensions of a non-densely defined $C$-symmetric contraction are studied and
parameterizations of all such extensions are obtained. As an application, a proof of an announced result of Glazman \cite{Glazman1} on the existence of maximal dissipative $C$-self-adjoint extensions of a densely defined $C$-symmetric dissipative operator is given. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 09:53:54 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Arlinskii",
"Yury",
""
],
[
"Schmüdgen",
"Konrad",
""
]
] |
2506.14358 | Lin-Ke Xie | Lin-Ke Xie, Wei Liu, Kaiyu Huang, Nai-Jie Guo, Jun-You Liu, Yu-Hang Ma, Ya-Qi Wu, Yi-Tao Wang, Zhao-an Wang, Xiao-Dong Zeng, Jia-Ming Ren, Chun Ao, Shuo Deng, Haifei Lu, Jian-Shun Tang, Chuan-Feng Li, Guang-Can Guo | Temperature dependent single- and double-quantum relaxation of negatively charged boron vacancies in hexagonal boron nitride | null | null | null | null | quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The negatively charged boron vacancy in two-dimensional hexagonal boron nitride has emerged as a promising candidate for quantum sensing. The coherence time of this defect spins which coherent quantum sensing resides in is limited spin-phonon interactions, while the underlying physical mechanism of the corresponding high-temperature behavior is still not fully understood. Here, we probe the single- and double-quantum relaxation rates on this center over the temperature range from 293 to 393 K. The results show that both relaxation rates increase with increasing temperature, and the double-quantum relaxation rate significantly increases rapidly. At high temperature (above 400 K), the double-quantum relaxation rate is much greater than single-quantum relaxation rate, and may dominate the decoherence channel of spin-phonon interactions. Using a theoretical model of second-order spin-phonon interactions, we attribute the high-temperature spin relaxation rates to interactions with higher-energy effective phonon mode, aiding the further understanding and guiding high-temperature sensing applications. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 09:58:21 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Xie",
"Lin-Ke",
""
],
[
"Liu",
"Wei",
""
],
[
"Huang",
"Kaiyu",
""
],
[
"Guo",
"Nai-Jie",
""
],
[
"Liu",
"Jun-You",
""
],
[
"Ma",
"Yu-Hang",
""
],
[
"Wu",
"Ya-Qi",
""
],
[
"Wang",
"Yi-Tao",
""
],
[
"Wang",
"Zhao-an",
""
],
[
"Zeng",
"Xiao-Dong",
""
],
[
"Ren",
"Jia-Ming",
""
],
[
"Ao",
"Chun",
""
],
[
"Deng",
"Shuo",
""
],
[
"Lu",
"Haifei",
""
],
[
"Tang",
"Jian-Shun",
""
],
[
"Li",
"Chuan-Feng",
""
],
[
"Guo",
"Guang-Can",
""
]
] |
2506.14359 | Sergej Monavari | Sergej Monavari | The refined local Donaldson-Thomas theory of curves | 50 pages, comments welcome | null | null | null | math.AG hep-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We solve the $K$-theoretically refined Donaldson-Thomas theory of local curves. Our results avoid degeneration techniques, but rather exploit direct localisation methods to reduce the refined Donaldson-Thomas partition function to the equivariant intersection theory of skew nested Hilbert schemes on smooth projective curves. We show that the latter is determined, for every Young diagram, by three universal series, which we compute in terms of the 1-leg $K$-theoretic equivariant vertex. In the refined limit, our results establish a formula for the refined topological string partition function of local curves proposed by Aganagic-Schaeffer. In the second part, we show that analogous structural results hold for the refined Pandharipande-Thomas theory of local curves. As an application, we deduce the K-theoretic DT/PT correspondence for local curves in arbitrary genus, as conjectured by Nekrasov-Okounkov.
Thanks to the recent machinery developed by Pardon, we expect our explicit results on local curves to play a key role towards the proof of the refined GW/PT conjectural correspondence of Brini-Schuler for all smooth Calabi-Yau threefolds. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 09:58:28 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Monavari",
"Sergej",
""
]
] |
2506.14360 | Yaning Zhao | Yaning Zhao, Luca Miszewski, Christian Deppe, Massimiliano Pierobon | Identification for Molecular Communication Based on Diffusion Channel with Poisson Reception Process | null | null | null | null | cs.IT math.IT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Molecular communication (MC) enables information exchange at the nano- and microscale, with applications in areas like drug delivery and health monitoring. These event-driven scenarios often require alternatives to traditional transmission. Identification communication, introduced by Ahlswede and Dueck, offers such an approach, in which the receiver only determines whether a specific message was sent, suiting resource-limited and event-triggered systems. This paper combines MC with identification and proposes a one-dimensional (1D) diffusion-based model. Diffusion noise is modeled as a Poisson process, and a lower bound on channel capacity is derived. Simulations, microscopic, and with short-length deterministic codes, validate theoretical results, including the channel impulse response and error bounds. The findings support the design of practical MC systems, with potential use in testbed development. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:00:31 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Zhao",
"Yaning",
""
],
[
"Miszewski",
"Luca",
""
],
[
"Deppe",
"Christian",
""
],
[
"Pierobon",
"Massimiliano",
""
]
] |
2506.14361 | Katherine Suess | Katherine A. Suess, Aliza G. Beverage, Mariska Kriek, Justin S. Spilker, Rachel Bezanson, Vincenzo R. D'Onofrio, Jenny E. Greene, Jamie Lin, Yuanze Luo, Desika Narayanan, Imad Pasha, Sedona H. Price, David J. Setton, Margaret E. Verrico, Yunchong Zhang | Cold gas in a post-starburst pair at z ~ 1.4: major mergers as a pathway to quenching in the HeavyMetal survey | 11 pages, 4 figures, submitted to ApJL. Comments welcome | null | null | null | astro-ph.GA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Recent observations at low redshift have revealed that some post-starburst galaxies retain significant molecular gas reservoirs despite low ongoing star formation rates, challenging theoretical predictions for galaxy quenching. To test whether this finding holds during the peak epoch of quenching, here we present ALMA CO(2-1) observations of five spectroscopically confirmed post-starburst galaxies at z ~ 1.4 from the HeavyMetal survey. While four galaxies are undetected in CO emission, we detect M_H2 ~ 10^9.7 Msun of molecular gas in one system. The detected system is a close pair of massive (M* = 10^(11.1-11.2) Msun) post-starburst galaxies with no clear tidal features, likely caught in the early stages of a major merger. These results suggest that mergers may be a key factor in retaining molecular gas while simultaneously suppressing star formation in quenched galaxies at high redshift, possibly by driving increased turbulence that decreases star formation efficiency. Unlike previous studies at z < 1, we find no correlation between molecular gas mass and time since quenching. This may be explained by the fact that -- despite having similar UVJ colors -- all galaxies in our sample have post-burst ages older than typical gas-rich quenched systems at low redshift. Our results highlight the importance of major mergers in shaping the cold gas content of quiescent galaxies during the peak epoch of quenching. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:01:48 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Suess",
"Katherine A.",
""
],
[
"Beverage",
"Aliza G.",
""
],
[
"Kriek",
"Mariska",
""
],
[
"Spilker",
"Justin S.",
""
],
[
"Bezanson",
"Rachel",
""
],
[
"D'Onofrio",
"Vincenzo R.",
""
],
[
"Greene",
"Jenny E.",
""
],
[
"Lin",
"Jamie",
""
],
[
"Luo",
"Yuanze",
""
],
[
"Narayanan",
"Desika",
""
],
[
"Pasha",
"Imad",
""
],
[
"Price",
"Sedona H.",
""
],
[
"Setton",
"David J.",
""
],
[
"Verrico",
"Margaret E.",
""
],
[
"Zhang",
"Yunchong",
""
]
] |
2506.14362 | Daniele Rege Cambrin | Daniele Rege Cambrin, Eleonora Poeta, Eliana Pastor, Isaac Corley, Tania Cerquitelli, Elena Baralis, Paolo Garza | HydroChronos: Forecasting Decades of Surface Water Change | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | Forecasting surface water dynamics is crucial for water resource management and climate change adaptation. However, the field lacks comprehensive datasets and standardized benchmarks. In this paper, we introduce HydroChronos, a large-scale, multi-modal spatiotemporal dataset for surface water dynamics forecasting designed to address this gap. We couple the dataset with three forecasting tasks. The dataset includes over three decades of aligned Landsat 5 and Sentinel-2 imagery, climate data, and Digital Elevation Models for diverse lakes and rivers across Europe, North America, and South America. We also propose AquaClimaTempo UNet, a novel spatiotemporal architecture with a dedicated climate data branch, as a strong benchmark baseline. Our model significantly outperforms a Persistence baseline for forecasting future water dynamics by +14% and +11% F1 across change detection and direction of change classification tasks, and by +0.1 MAE on the magnitude of change regression. Finally, we conduct an Explainable AI analysis to identify the key climate variables and input channels that influence surface water change, providing insights to inform and guide future modeling efforts. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:02:48 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Cambrin",
"Daniele Rege",
""
],
[
"Poeta",
"Eleonora",
""
],
[
"Pastor",
"Eliana",
""
],
[
"Corley",
"Isaac",
""
],
[
"Cerquitelli",
"Tania",
""
],
[
"Baralis",
"Elena",
""
],
[
"Garza",
"Paolo",
""
]
] |
2506.14363 | Oliver Markgraf | Matthew Hague, Denghang Hu, Artur Je\.z, Anthony W. Lin, Oliver Markgraf, Philipp R\"ummer, Zhilin Wu | OSTRICH2: Solver for Complex String Constraints | null | null | null | null | cs.LO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present OSTRICH2, the latest evolution of the SMT solver OSTRICH for string constraints. OSTRICH2 supports a wide range of complex functions on strings and provides completeness guarantees for a substantial fragment of string constraints, including the straight-line fragment and the chain-free fragment. OSTRICH2 provides full support for the SMT-LIB theory of Unicode strings, extending the standard with several unique features not found in other solvers: among others, parsing of ECMAScript regular expressions (including look-around assertions and capture groups) and handling of user-defined string transducers. We empirically demonstrate that OSTRICH2 is competitive to other string solvers on SMT-COMP benchmarks. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:03:37 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Hague",
"Matthew",
""
],
[
"Hu",
"Denghang",
""
],
[
"Jeż",
"Artur",
""
],
[
"Lin",
"Anthony W.",
""
],
[
"Markgraf",
"Oliver",
""
],
[
"Rümmer",
"Philipp",
""
],
[
"Wu",
"Zhilin",
""
]
] |
2506.14364 | Weiyu Zhou | Weiyu Zhou, Zheng Wang, Chao Chen, Yike Li, Yongkui Yang, Zhuoyu Wu, Anupam Chattopadhyay | Tensor Manipulation Unit (TMU): Reconfigurable, Near-Memory Tensor Manipulation for High-Throughput AI SoC | 10 pages | null | null | null | cs.AR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | While recent advances in AI SoC design have focused heavily on accelerating tensor computation, the equally critical task of tensor manipulation, centered on high,volume data movement with minimal computation, remains underexplored. This work addresses that gap by introducing the Tensor Manipulation Unit (TMU), a reconfigurable, near-memory hardware block designed to efficiently execute data-movement-intensive operators. TMU manipulates long datastreams in a memory-to-memory fashion using a RISC-inspired execution model and a unified addressing abstraction, enabling broad support for both coarse- and fine-grained tensor transformations. Integrated alongside a TPU within a high-throughput AI SoC, the TMU leverages double buffering and output forwarding to improve pipeline utilization. Fabricated in SMIC 40nm technology, the TMU occupies only 0.019 mm2 while supporting over 10 representative tensor manipulation operators. Benchmarking shows that TMU alone achieves up to 1413 and 8.54 operator-level latency reduction compared to ARM A72 and NVIDIA Jetson TX2, respectively. When integrated with the in-house TPU, the complete system achieves a 34.6% reduction in end-to-end inference latency, demonstrating the effectiveness and scalability of reconfigurable tensor manipulation in modern AI SoCs. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:05:44 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Zhou",
"Weiyu",
""
],
[
"Wang",
"Zheng",
""
],
[
"Chen",
"Chao",
""
],
[
"Li",
"Yike",
""
],
[
"Yang",
"Yongkui",
""
],
[
"Wu",
"Zhuoyu",
""
],
[
"Chattopadhyay",
"Anupam",
""
]
] |
2506.14365 | Wim Vanderbauwhede | Wim Vanderbauwhede, Mattias Wadenstein | Life Cycle Analysis for Emissions of Scientific Computing Centres | null | null | null | null | hep-ex | http://creativecommons.org/licenses/by-sa/4.0/ | We propose a dedicated model to assist with the life cycle analysis of emissions of scientific computing centres. The model takes into account both the embodied carbon and emissions from use, as well as other factors such as data centre power usage efficiency, data centre expansion, hardware replacement, increase in energy efficiency of next-generation hardware, reduction in carbon intensity of the electricity supply and potential for heat reuse. If differs from existing models in its detailed handling of hardware embodied carbon and time dependency of various factors affecting the emissions. We present a number of scenarios where we apply the model to real-life HPC centres in different countries to illustrate how the trade-offs depend on the various factors and validate our model against the literature. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:06:25 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Vanderbauwhede",
"Wim",
""
],
[
"Wadenstein",
"Mattias",
""
]
] |
2506.14366 | Jing Liu | Chengjie Fu, Jing Liu | The power of SKA to Constrain cosmological gravitational-wave backgrounds below the astrophysical foreground noise | 8 pages, 2 figures | null | null | null | astro-ph.CO gr-qc hep-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The inspirals of supermaissive black hole binaries provide a convinced gravitational wave background in the nHz band, serving as the fiducial model of the recent gravitational wave signal reported by the PTA experiments. The uncertainties of the number of binaries contributing to each frequency bin introduce a foreground noise in the nHz and $\mu$Hz bands against the observation of the underlying gravitational wave backgrounds of the cosmological origin. In this work, we investigate a new method to constrain the cosmological gravitational wave strength under the astrophysical foreground. The energy density fluctuations from cosmological gravitational-wave sources can generally trigger the formation of compact subhalos of dark matter, and the upcoming Square Kilometer Array has the ability to constrain the abundance of the subhalos at the $\mathcal{O}(1)$ level. The cosmological gravitational wave energy spectra from various sources are expected to be constrained several orders of magnitude below the astrophysical foreground, providing more strict constraints on the parameter spaces of corresponding new physics models. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:07:51 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Fu",
"Chengjie",
""
],
[
"Liu",
"Jing",
""
]
] |
2506.14367 | Mahin Montasir Afif | Sumshun Nahar Eity, Mahin Montasir Afif, Tanisha Fairooz, Md. Mortuza Ahmmed, Md Saef Ullah Miah | DGG-XNet: A Hybrid Deep Learning Framework for Multi-Class Brain Disease Classification with Explainable AI | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurate diagnosis of brain disorders such as Alzheimer's disease and brain tumors remains a critical challenge in medical imaging. Conventional methods based on manual MRI analysis are often inefficient and error-prone. To address this, we propose DGG-XNet, a hybrid deep learning model integrating VGG16 and DenseNet121 to enhance feature extraction and classification. DenseNet121 promotes feature reuse and efficient gradient flow through dense connectivity, while VGG16 contributes strong hierarchical spatial representations. Their fusion enables robust multiclass classification of neurological conditions. Grad-CAM is applied to visualize salient regions, enhancing model transparency. Trained on a combined dataset from BraTS 2021 and Kaggle, DGG-XNet achieved a test accuracy of 91.33\%, with precision, recall, and F1-score all exceeding 91\%. These results highlight DGG-XNet's potential as an effective and interpretable tool for computer-aided diagnosis (CAD) of neurodegenerative and oncological brain disorders. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:07:59 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Eity",
"Sumshun Nahar",
""
],
[
"Afif",
"Mahin Montasir",
""
],
[
"Fairooz",
"Tanisha",
""
],
[
"Ahmmed",
"Md. Mortuza",
""
],
[
"Miah",
"Md Saef Ullah",
""
]
] |
2506.14368 | Ashish Kumar | Ashish Kumar, Adam T. Deller, Pankaj Jain, and Javier Mold\'on | VLBA astrometry of PSRs B0329+54 and B1133+16: Improved pulsar distances and comparison of global ionospheric models | Accepted for publication in PASA | null | null | null | astro-ph.HE astro-ph.GA astro-ph.IM astro-ph.SR | http://creativecommons.org/licenses/by/4.0/ | Very long baseline interferometry (VLBI) astrometry is used to determine the three-dimensional position and proper motion of astronomical objects. A typical VLBI astrometric campaign generally includes around ten observations, making it challenging to characterise systematic uncertainties. Our study on two bright pulsars, B0329+54 and B1133+16, involves analysis of broadband Very Long Baseline Array (VLBA) data over $\sim30$ epochs (spanning approximately $3.5\, {\rm years}$). This extended dataset has significantly improved the precision of the astrometric estimates of these pulsars. Our broadband study suggests that, as expected, the primary contribution to systematic uncertainties in L-band VLBI astrometry originates from the ionosphere. We have also assessed the effectiveness of the modified TEC (total electron content) mapping function, which converts vertical TEC to slant TEC, in correcting ionospheric dispersive delays using global TEC maps. The parallax and proper motion obtained from the multiple data sets, calibrated using the traditional and the modified TEC mapping functions, are consistent. However, the reduced chi-square values from least-squares fitting and precision of the fitted astrometric parameters show no significant improvement, and hence, the effectiveness of the new TEC mapping function on astrometry is unclear. For B0329+54, the refined parallax estimate is $0.611^{+0.013}_{-0.013}\, {\rm mas}$, with best-fit proper motion of $\mu_{\alpha} = 16.960^{+0.011}_{-0.010}\, {\rm mas\, yr^{-1}}$ in R.A. and and $\mu_{\delta} = -10.382^{+0.022}_{-0.022}\, {\rm mas\, yr^{-1}}$ in Dec. For B1133+16, the new estimated parallax is $2.705^{+0.009}_{-0.009}\, {\rm mas}$, with proper motions of $\mu_{\alpha} = -73.777^{+0.008}_{-0.008}\, {\rm mas\, yr^{-1}}$ and $\mu_{\delta} = 366.573^{+0.019}_{-0.019}\, {\rm mas\, yr^{-1}}$. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:08:58 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Kumar",
"Ashish",
""
],
[
"Deller",
"Adam T.",
""
],
[
"Jain",
"Pankaj",
""
],
[
"Moldón",
"Javier",
""
]
] |
2506.14369 | Maria Spichkova | Maria Spichkova | Agile and Student-Centred Teaching of Agile/Scrum Concepts | Preprint. Accepted to the 29th International Conference on Knowledge-Based and Intelligent Information & Engineering Systems (KES 2025). Final version to be published by Elsevier (In Press) | null | null | null | cs.SE | http://creativecommons.org/licenses/by-nc-nd/4.0/ | In this paper, we discuss our experience in designing and teaching a course on Software Engineering Project Management, where the focus is on Agile/Scrum development and Requirement Engineering activities. The course has undergone fundamental changes since 2020 to make the teaching approach more student-centred and flexible. As many universities abandoned having face-to-face exams at the end of the semester, authentic assessments now play an even more important role than before. This makes assessment of students' work even more challenging, especially if we are dealing with large cohorts of students. The complexity is not only in dealing with diversity in the student cohorts when elaborating the assessment tasks, but also in being able to provide feedback and marks in a timely and fairly. We report our lessons learned, which might provide useful insights for teaching Agile/Scrum concepts to undergraduate and postgraduate students. We also analyse what course structure might be effective to support a blended learning approach, as well as what could be a reasonable structure of online assessments, to keep them both authentic and scalable for large cohorts of students. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:09:32 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Spichkova",
"Maria",
""
]
] |
2506.14370 | Amrit Poudel | Amrit Poudel, Yifan Ding, Jurgen Pfeffer, Tim Weninger | Digital Gatekeepers: Google's Role in Curating Hashtags and Subreddits | Accepted to ACL 2025 Main | Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics 2025 | null | null | cs.CL | http://creativecommons.org/licenses/by/4.0/ | Search engines play a crucial role as digital gatekeepers, shaping the visibility of Web and social media content through algorithmic curation. This study investigates how search engines like Google selectively promotes or suppresses certain hashtags and subreddits, impacting the information users encounter. By comparing search engine results with nonsampled data from Reddit and Twitter/X, we reveal systematic biases in content visibility. Google's algorithms tend to suppress subreddits and hashtags related to sexually explicit material, conspiracy theories, advertisements, and cryptocurrencies, while promoting content associated with higher engagement. These findings suggest that Google's gatekeeping practices influence public discourse by curating the social media narratives available to users. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:10:39 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Poudel",
"Amrit",
""
],
[
"Ding",
"Yifan",
""
],
[
"Pfeffer",
"Jurgen",
""
],
[
"Weninger",
"Tim",
""
]
] |
2506.14371 | Lucile Favero | Lucile Favero, Daniel Frases, Juan Antonio P\'erez-Ortiz, Tanja K\"aser, and Nuria Oliver | ELLIS Alicante at CQs-Gen 2025: Winning the critical thinking questions shared task: LLM-based question generation and selection | Proceedings of the 12th Workshop on Argument Mining | null | null | null | cs.CL cs.HC | http://creativecommons.org/licenses/by/4.0/ | The widespread adoption of chat interfaces based on Large Language Models (LLMs) raises concerns about promoting superficial learning and undermining the development of critical thinking skills. Instead of relying on LLMs purely for retrieving factual information, this work explores their potential to foster deeper reasoning by generating critical questions that challenge unsupported or vague claims in debate interventions. This study is part of a shared task of the 12th Workshop on Argument Mining, co-located with ACL 2025, focused on automatic critical question generation. We propose a two-step framework involving two small-scale open source language models: a Questioner that generates multiple candidate questions and a Judge that selects the most relevant ones. Our system ranked first in the shared task competition, demonstrating the potential of the proposed LLM-based approach to encourage critical engagement with argumentative texts. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:10:51 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Favero",
"Lucile",
""
],
[
"Frases",
"Daniel",
""
],
[
"Pérez-Ortiz",
"Juan Antonio",
""
],
[
"Käser",
"Tanja",
""
],
[
"Oliver",
"Nuria",
""
]
] |
2506.14372 | Radmehr Jafari | Sasan Kheiri, R. Jafari, S. Mahdavifar, Ehsan Nedaaee Oskoee, and Alireza Akbari | Dynamical Phase diagram of the Quantum Ising model with Cluster Interaction Under Noisy and Noiseless Driven field | null | null | null | null | cond-mat.stat-mech quant-ph | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In most lattice models, gap closing typically occurs at high-symmetry points in the Brillouin zone. In the transverse field Ising model with cluster interaction, besides the gap closing at high-symmetry points, the gap closing at the quantum phase transition between paramagnetic and cluster phases of the model can be moved by tuning the strength of the cluster interaction. We take advantage of this property to examine the nonequilibrium dynamics of the model in the framework of dynamical quantum phase transitions (DQPTs) after a noiseless and noisy ramp of the transverse magnetic field. The numerical results show that DQPTs always happen if the starting or ending point of the quench field is restricted between two critical points. In other ways, there is always critical sweep velocity above which DQPTs disappear. Our finding reveals that noise modifies drastically the dynamical phase diagram of the model. We find that the critical sweep velocity decreases by enhancing the noise intensity and scales linearly with the square of noise intensity for weak and strong noise. Moreover, the region with multi-critical modes induced in the dynamical phase diagram by noise. The sweep velocity under which the system enters the multi-critical modes (MCMs) region increases by enhancing the noise and scales linearly with the square of noise intensity | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:13:43 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Kheiri",
"Sasan",
""
],
[
"Jafari",
"R.",
""
],
[
"Mahdavifar",
"S.",
""
],
[
"Oskoee",
"Ehsan Nedaaee",
""
],
[
"Akbari",
"Alireza",
""
]
] |
2506.14373 | Junyeob Baek | Junyeob Baek, Hosung Lee, Christopher Hoang, Mengye Ren, Sungjin Ahn | Discrete JEPA: Learning Discrete Token Representations without Reconstruction | null | null | null | null | cs.CV | http://creativecommons.org/licenses/by/4.0/ | The cornerstone of cognitive intelligence lies in extracting hidden patterns from observations and leveraging these principles to systematically predict future outcomes. However, current image tokenization methods demonstrate significant limitations in tasks requiring symbolic abstraction and logical reasoning capabilities essential for systematic inference. To address this challenge, we propose Discrete-JEPA, extending the latent predictive coding framework with semantic tokenization and novel complementary objectives to create robust tokenization for symbolic reasoning tasks. Discrete-JEPA dramatically outperforms baselines on visual symbolic prediction tasks, while striking visual evidence reveals the spontaneous emergence of deliberate systematic patterns within the learned semantic token space. Though an initial model, our approach promises a significant impact for advancing Symbolic world modeling and planning capabilities in artificial intelligence systems. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:15:17 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Baek",
"Junyeob",
""
],
[
"Lee",
"Hosung",
""
],
[
"Hoang",
"Christopher",
""
],
[
"Ren",
"Mengye",
""
],
[
"Ahn",
"Sungjin",
""
]
] |
2506.14374 | Wai Man Si | Wai Man Si, Mingjie Li, Michael Backes, Yang Zhang | Excessive Reasoning Attack on Reasoning LLMs | null | null | null | null | cs.CR cs.LG | http://creativecommons.org/licenses/by/4.0/ | Recent reasoning large language models (LLMs), such as OpenAI o1 and DeepSeek-R1, exhibit strong performance on complex tasks through test-time inference scaling. However, prior studies have shown that these models often incur significant computational costs due to excessive reasoning, such as frequent switching between reasoning trajectories (e.g., underthinking) or redundant reasoning on simple questions (e.g., overthinking). In this work, we expose a novel threat: adversarial inputs can be crafted to exploit excessive reasoning behaviors and substantially increase computational overhead without compromising model utility. Therefore, we propose a novel loss framework consisting of three components: (1) Priority Cross-Entropy Loss, a modification of the standard cross-entropy objective that emphasizes key tokens by leveraging the autoregressive nature of LMs; (2) Excessive Reasoning Loss, which encourages the model to initiate additional reasoning paths during inference; and (3) Delayed Termination Loss, which is designed to extend the reasoning process and defer the generation of final outputs. We optimize and evaluate our attack for the GSM8K and ORCA datasets on DeepSeek-R1-Distill-LLaMA and DeepSeek-R1-Distill-Qwen. Empirical results demonstrate a 3x to 9x increase in reasoning length with comparable utility performance. Furthermore, our crafted adversarial inputs exhibit transferability, inducing computational overhead in o3-mini, o1-mini, DeepSeek-R1, and QWQ models. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:16:52 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Si",
"Wai Man",
""
],
[
"Li",
"Mingjie",
""
],
[
"Backes",
"Michael",
""
],
[
"Zhang",
"Yang",
""
]
] |
2506.14375 | Muhammad Hamza Yousuf | Muhammad Hamza Yousuf, Jason Li, Sahar Vahdati, Raphael Theilen, Jakob Wittenstein, Jens Lehmann | IntelliLung: Advancing Safe Mechanical Ventilation using Offline RL with Hybrid Actions and Clinically Aligned Rewards | under review, PAIS track @ ECAI 2025 | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Invasive mechanical ventilation (MV) is a life-sustaining therapy for critically ill patients in the intensive care unit (ICU). However, optimizing its settings remains a complex and error-prone process due to patient-specific variability. While Offline Reinforcement Learning (RL) shows promise for MV control, current stateof-the-art (SOTA) methods struggle with the hybrid (continuous and discrete) nature of MV actions. Discretizing the action space limits available actions due to exponential growth in combinations and introduces distribution shifts that can compromise safety. In this paper, we propose optimizations that build upon prior work in action space reduction to address the challenges of discrete action spaces. We also adapt SOTA offline RL algorithms (IQL and EDAC) to operate directly on hybrid action spaces, thereby avoiding the pitfalls of discretization. Additionally, we introduce a clinically grounded reward function based on ventilator-free days and physiological targets, which provides a more meaningful optimization objective compared to traditional sparse mortality-based rewards. Our findings demonstrate that AI-assisted MV optimization may enhance patient safety and enable individualized lung support, representing a significant advancement toward intelligent, data-driven critical care solutions. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:17:26 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Yousuf",
"Muhammad Hamza",
""
],
[
"Li",
"Jason",
""
],
[
"Vahdati",
"Sahar",
""
],
[
"Theilen",
"Raphael",
""
],
[
"Wittenstein",
"Jakob",
""
],
[
"Lehmann",
"Jens",
""
]
] |
2506.14376 | Giuseppe Riva | Massimo Chiriatti, Marianna Bergamaschi Ganapini, Enrico Panai, Brenda K. Wiederhold, Giuseppe Riva | System 0: Transforming Artificial Intelligence into a Cognitive Extension | null | null | 10.1089/cyber.2025.0201 | null | cs.HC | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This paper introduces System 0, a conceptual framework for understanding how artificial intelligence functions as a cognitive extension preceding both intuitive (System 1) and deliberative (System 2) thinking processes. As AI systems increasingly shape the informational substrate upon which human cognition operates, they transform from passive tools into active cognitive partners. Building on the Extended Mind hypothesis and Heersmink's criteria for cognitive extension, we argue that AI systems satisfy key conditions for cognitive integration. These include reliability, trust, transparency, individualization, and the ability to enhance and transform human mental functions. However, AI integration creates a paradox: while expanding cognitive capabilities, it may simultaneously constrain thinking through sycophancy and bias amplification. To address these challenges, we propose seven evidence-based frameworks for effective human-AI cognitive integration: Enhanced Cognitive Scaffolding, which promotes progressive autonomy; Symbiotic Division of Cognitive Labor, strategically allocating tasks based on comparative strengths; Dialectical Cognitive Enhancement, countering AI sycophancy through productive epistemic tension; Agentic Transparency and Control, ensuring users understand and direct AI influence; Expertise Democratization, breaking down knowledge silos; Social-Emotional Augmentation, addressing affective dimensions of cognitive work; and Duration-Optimized Integration, managing the evolving human-AI relationship over time. Together, these frameworks provide a comprehensive approach for harnessing AI as a genuine cognitive extension while preserving human agency, critical thinking, and intellectual growth, transforming AI from a replacement for human cognition into a catalyst for enhanced thinking. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:19:46 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Chiriatti",
"Massimo",
""
],
[
"Ganapini",
"Marianna Bergamaschi",
""
],
[
"Panai",
"Enrico",
""
],
[
"Wiederhold",
"Brenda K.",
""
],
[
"Riva",
"Giuseppe",
""
]
] |
2506.14377 | Qin Xie | Qin Xie, Ming Li, Fei Cheng | Between Regulation and Accessibility: How Chinese University Students Navigate Global and Domestic Generative AI | null | null | null | null | cs.CY | http://creativecommons.org/licenses/by/4.0/ | Despite the rapid proliferation of generative AI in higher education, students in China face significant barriers in accessing global tools like ChatGPT due to regulations and constraints. Grounded in the Unified Theory of Acceptance and Use of Technology 2 (UTAUT2) model, this study employs qualitative interviews to investigate how Chinese university students interact with both global and domestic generative AIs in the learning process. Findings reveal that engagement is shaped by accessibility, language proficiency, and cultural relevance. Students often employ workarounds (e.g., VPNs) to access global generative AIs, raising ethical and privacy concerns. Domestic generative AIs, while offering language and cultural advantages, are limited by content filtering and output constraints. This research contributes to understanding generative AI adoption in non-Western contexts by highlighting the complex interplay of political, linguistic, and cultural factors. It advocates for human-centered, multilingual, domestic context-sensitive AI integration to ensure equitable and inclusive digital learning environments. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:24:12 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Xie",
"Qin",
""
],
[
"Li",
"Ming",
""
],
[
"Cheng",
"Fei",
""
]
] |
2506.14378 | G.Cigdem Yalcin | G.Cigdem Yalcin and M.Berk Onder | A novel approach for converting spatio-temporal series into complex networks | 6 pages | null | null | null | physics.soc-ph cond-mat.dis-nn cond-mat.stat-mech | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | This study aims to offer a new perspective on complex network representation of real-world systems. Currently, the most well-known transformation algorithms in the literature treat each data point in a time series as a node and transform the time series into a network. In this study, we present a new approach converting spatio-temporal series into a complex network. We focus on studying this transformation by grounding it in the context of physics , with the aim of adapting it to real-world problems, which often manifest as complex systems across various domains. We introduce the Gravitational Graph (GG) algorithm, which is grounded in the concept of gravitational force from fundamental physics. We consider air pollution concentrations, which represent a global environmental health risk, as an example of a complex environmental system, and apply the GG algorithm to particulate matter of 10 microns (PM10) recorded by 21 air quality monitoring stations located in various regions of Istanbul, Turkiye. While the GG algorithm allows for the conversion of spatio-temporal series -- rather than time series -- into networks, it also enables the analysis of the statistical properties and characteristics of the converted networks, thereby uncovering hidden relationships and dependencies that may not be apparent in the original time series. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:24:19 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Yalcin",
"G. Cigdem",
""
],
[
"Onder",
"M. Berk",
""
]
] |
2506.14379 | Seyran Ibrahimov | Seyran S. Ibrahimov, Nazim I. Mahmudov | On Certain Diophantine Equations Involving Lucas and Pell Numbers | null | null | null | null | math.NT | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Let \( L_t \) denote the \( t \)-th term of the Lucas sequence and \( P_t \) the \( t \)-th term of the Pell sequence. In this paper, we prove that the Diophantine equation \[ L_m^{n+k} + L_m^n = L_r \] has no solution in positive integers \( r, m, n, k \) with \( m, n \geq 2 \). Additionally, we show that \( (r, m, n, k) = (4, 2, 2, 1) \) is the only solution in positive integers to the Diophantine equation \[ P_m^{n+k} + P_m^n = P_r \] under the same condition \( m, n \geq 2 \). | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:24:32 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Ibrahimov",
"Seyran S.",
""
],
[
"Mahmudov",
"Nazim I.",
""
]
] |
2506.14380 | Herve Bouy | H. Bouy, G. Duch\^ene, G. Strampelli, J. Aguilar, J. Olivares, D. Barrado, S. N. Raymond, N. Hu\'elamo, M. Tamura, E. Bertin, W. Brandner, J.-C. Cuillandre, P. A. B. Galli, and N. Miret-Roig | Multiplicity of young isolated planetary mass objects in Taurus and Upper Scorpius | Submitted to A&A on 2025-04-04 ; awaiting the initial review | null | null | null | astro-ph.SR astro-ph.EP astro-ph.GA | http://creativecommons.org/licenses/by/4.0/ | Free-floating planetary mass objects-worlds that roam interstellar space untethered to a parent star-challenge conventional notions of planetary formation and migration, but also of star and brown dwarf formation. We focus on the multiplicity among free-floating planets. By virtue of their low binding energy (compared to other objects formed in these environments), these low-mass substellar binaries represent a most sensitive probe of the mechanisms at play during the star formation process. We use the Hubble Space Telescope and its Wide Field Camera 3 and the Very Large Telescope and its ERIS adaptive optics facility to search for visual companions among a sample of 77 objects members of the Upper Scorpius and Taurus young nearby associations with estimated masses in the range between approximately 5-33 M$_{\rm Jup}$. We report the discovery of one companion candidate around a Taurus member with a separation of 111.9$\pm$0.4 mas, or $\sim$18 au assuming a distance of 160pc, with an estimated primary mass in the range between 3-6 M$_{\rm Jup}$ and a secondary mass between 2.6-5.2 M$_{\rm Jup}$, depending on the assumed age. This corresponds to an overall binary fraction of 1.8$^{+2.6}_{-1.3}$% among free-floating planetary mass objects over the separation range $\ge$7 au. Despite the limitations of small-number statistics and variations in spatial resolution and sensitivity, our results, combined with previous high-spatial-resolution surveys, suggest a notable difference in the multiplicity properties of objects below $\sim$25 M$_{\rm Jup}$ between Upper Sco and Taurus. In Taurus, five companions were identified among 78 observed objects (4.9$^{+2.8}_{-2.0}$%), whereas none were found among 97 objects in Upper Sco ($\le$1.2%).} | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:24:36 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Bouy",
"H.",
""
],
[
"Duchêne",
"G.",
""
],
[
"Strampelli",
"G.",
""
],
[
"Aguilar",
"J.",
""
],
[
"Olivares",
"J.",
""
],
[
"Barrado",
"D.",
""
],
[
"Raymond",
"S. N.",
""
],
[
"Huélamo",
"N.",
""
],
[
"Tamura",
"M.",
""
],
[
"Bertin",
"E.",
""
],
[
"Brandner",
"W.",
""
],
[
"Cuillandre",
"J. -C.",
""
],
[
"Galli",
"P. A. B.",
""
],
[
"Miret-Roig",
"N.",
""
]
] |
2506.14381 | Yuxuan Jiang | Yuxuan Jiang, Siyue Teng, Qiang Zhu, Chen Feng, Chengxi Zeng, Fan Zhang, Shuyuan Zhu, Bing Zeng, David Bull | Compressed Video Super-Resolution based on Hierarchical Encoding | null | null | null | null | eess.IV cs.CV | http://creativecommons.org/licenses/by-nc-nd/4.0/ | This paper presents a general-purpose video super-resolution (VSR) method, dubbed VSR-HE, specifically designed to enhance the perceptual quality of compressed content. Targeting scenarios characterized by heavy compression, the method upscales low-resolution videos by a ratio of four, from 180p to 720p or from 270p to 1080p. VSR-HE adopts hierarchical encoding transformer blocks and has been sophisticatedly optimized to eliminate a wide range of compression artifacts commonly introduced by H.265/HEVC encoding across various quantization parameter (QP) levels. To ensure robustness and generalization, the model is trained and evaluated under diverse compression settings, allowing it to effectively restore fine-grained details and preserve visual fidelity. The proposed VSR-HE has been officially submitted to the ICME 2025 Grand Challenge on VSR for Video Conferencing (Team BVI-VSR), under both the Track 1 (General-Purpose Real-World Video Content) and Track 2 (Talking Head Videos). | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:26:07 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Jiang",
"Yuxuan",
""
],
[
"Teng",
"Siyue",
""
],
[
"Zhu",
"Qiang",
""
],
[
"Feng",
"Chen",
""
],
[
"Zeng",
"Chengxi",
""
],
[
"Zhang",
"Fan",
""
],
[
"Zhu",
"Shuyuan",
""
],
[
"Zeng",
"Bing",
""
],
[
"Bull",
"David",
""
]
] |
2506.14382 | Ning Zhou | Ning Zhou, Shanxiong Chen, Mingting Zhou, Haigang Sui, Lieyun Hu, Han Li, Li Hua, Qiming Zhou | DepthSeg: Depth prompting in remote sensing semantic segmentation | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Remote sensing semantic segmentation is crucial for extracting detailed land surface information, enabling applications such as environmental monitoring, land use planning, and resource assessment. In recent years, advancements in artificial intelligence have spurred the development of automatic remote sensing semantic segmentation methods. However, the existing semantic segmentation methods focus on distinguishing spectral characteristics of different objects while ignoring the differences in the elevation of the different targets. This results in land cover misclassification in complex scenarios involving shadow occlusion and spectral confusion. In this paper, we introduce a depth prompting two-dimensional (2D) remote sensing semantic segmentation framework (DepthSeg). It automatically models depth/height information from 2D remote sensing images and integrates it into the semantic segmentation framework to mitigate the effects of spectral confusion and shadow occlusion. During the feature extraction phase of DepthSeg, we introduce a lightweight adapter to enable cost-effective fine-tuning of the large-parameter vision transformer encoder pre-trained by natural images. In the depth prompting phase, we propose a depth prompter to model depth/height features explicitly. In the semantic prediction phase, we introduce a semantic classification decoder that couples the depth prompts with high-dimensional land-cover features, enabling accurate extraction of land-cover types. Experiments on the LiuZhou dataset validate the advantages of the DepthSeg framework in land cover mapping tasks. Detailed ablation studies further highlight the significance of the depth prompts in remote sensing semantic segmentation. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:27:59 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Zhou",
"Ning",
""
],
[
"Chen",
"Shanxiong",
""
],
[
"Zhou",
"Mingting",
""
],
[
"Sui",
"Haigang",
""
],
[
"Hu",
"Lieyun",
""
],
[
"Li",
"Han",
""
],
[
"Hua",
"Li",
""
],
[
"Zhou",
"Qiming",
""
]
] |
2506.14383 | Sabri Koraltan | Sabri Koraltan, Fabrizio Porrati, Robert Kraft, Sven Barth, Markus Weigand, Claas Abert, Dieter Suess, Michael Huth, Sebastian Wintz | Reconfigurable three dimensional magnetic nanoarchitectures | 10 pages, 6 figures | null | null | null | cond-mat.mes-hall cond-mat.mtrl-sci physics.comp-ph | http://creativecommons.org/licenses/by/4.0/ | Three-dimensional (3D) nanomagnetism is a rapidly developing field within magnetic materials research, where exploiting the third dimension unlocks opportunities for innovative applications in areas such as sensing, data storage, and neuromorphic computing. Among various fabrication techniques, focused electron beam-induced deposition (FEBID) offers high flexibility in creating complex 3D nanostructures with sub-100 nm resolution. A key challenge in the development of 3D nanomagnets is the ability to locally control the magnetic configuration, which is essential to achieve desired functionalities. In this work, the magnetization reversal mechanism of a three-dimensional nanoarchitecture fabricated using focused electron beam-induced deposition is investigated by combining direct observation via scanning transmission X-ray microscopy with finite element micromagnetic simulations. In particular, our investigation shows that the magnetization of the components of a three-dimensional Co3 Fe tetrapod can be reversed individually and sequentially. Finally, it is demonstrated that complete control and reconfigurability of the system can be achieved by tuning the direction of the applied magnetic field. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:30:13 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Koraltan",
"Sabri",
""
],
[
"Porrati",
"Fabrizio",
""
],
[
"Kraft",
"Robert",
""
],
[
"Barth",
"Sven",
""
],
[
"Weigand",
"Markus",
""
],
[
"Abert",
"Claas",
""
],
[
"Suess",
"Dieter",
""
],
[
"Huth",
"Michael",
""
],
[
"Wintz",
"Sebastian",
""
]
] |
2506.14384 | Huan Kang | Huan Kang, Hui Li, Xiao-Jun Wu, Tianyang Xu, Rui Wang, Chunyang Cheng, Josef Kittler | GrFormer: A Novel Transformer on Grassmann Manifold for Infrared and Visible Image Fusion | 16 pages, 11 figures | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the field of image fusion, promising progress has been made by modeling data from different modalities as linear subspaces.
However, in practice, the source images are often located in a non-Euclidean space, where the Euclidean methods usually cannot
encapsulate the intrinsic topological structure. Typically, the inner product performed in the Euclidean space calculates the algebraic
similarity rather than the semantic similarity, which results in undesired attention output and a decrease in fusion performance.
While the balance of low-level details and high-level semantics should be considered in infrared and visible image fusion task. To
address this issue, in this paper, we propose a novel attention mechanism based on Grassmann manifold for infrared and visible
image fusion (GrFormer). Specifically, our method constructs a low-rank subspace mapping through projection constraints on the
Grassmann manifold, compressing attention features into subspaces of varying rank levels. This forces the features to decouple into
high-frequency details (local low-rank) and low-frequency semantics (global low-rank), thereby achieving multi-scale semantic
fusion. Additionally, to effectively integrate the significant information, we develop a cross-modal fusion strategy (CMS) based on
a covariance mask to maximise the complementary properties between different modalities and to suppress the features with high
correlation, which are deemed redundant. The experimental results demonstrate that our network outperforms SOTA methods both
qualitatively and quantitatively on multiple image fusion benchmarks. The codes are available at https://github.com/Shaoyun2023. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:32:05 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Kang",
"Huan",
""
],
[
"Li",
"Hui",
""
],
[
"Wu",
"Xiao-Jun",
""
],
[
"Xu",
"Tianyang",
""
],
[
"Wang",
"Rui",
""
],
[
"Cheng",
"Chunyang",
""
],
[
"Kittler",
"Josef",
""
]
] |
2506.14385 | Amy Siobhan Inwood | Amy S. Inwood, Peter J. Smith, Mahmoud AlaaEldin, Michail Matthaiou | Performance Characterization of Continuous Reconfigurable Intelligent Surfaces | null | null | null | null | eess.SP | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We consider a reconfigurable intelligent surface (RIS) that can implement a phase rotation continuously over the whole surface rather than via a finite number of discrete elements. Such an RIS can be considered a design for future systems where advances in metamaterials make such an implementation feasible or as the limiting case where the number of elements in a traditional RIS increases in a given area. We derive the optimal RIS design for the single-user (SU) scenario assuming a line-of-sight (LoS) from the RIS to the base station (BS) and correlated Rayleigh fading for the other links. We also derive the associated optimal signal-to-noise ratio (SNR) and its mean, a bound on the mean spectral efficiency (SE), an approximation to the SNR outage probability and an approximation to the coefficient of variation for the investigation of channel hardening. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:32:18 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Inwood",
"Amy S.",
""
],
[
"Smith",
"Peter J.",
""
],
[
"AlaaEldin",
"Mahmoud",
""
],
[
"Matthaiou",
"Michail",
""
]
] |
2506.14386 | Christian H.X. Ali Mehmeti-G\"opel | Christian H.X. Ali Mehmeti-G\"opel, Michael Wand | ResNets Are Deeper Than You Think | NeurIPS 2025 Submission | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Residual connections remain ubiquitous in modern neural network architectures nearly a decade after their introduction. Their widespread adoption is often credited to their dramatically improved trainability: residual networks train faster, more stably, and achieve higher accuracy than their feedforward counterparts. While numerous techniques, ranging from improved initialization to advanced learning rate schedules, have been proposed to close the performance gap between residual and feedforward networks, this gap has persisted. In this work, we propose an alternative explanation: residual networks do not merely reparameterize feedforward networks, but instead inhabit a different function space. We design a controlled post-training comparison to isolate generalization performance from trainability; we find that variable-depth architectures, similar to ResNets, consistently outperform fixed-depth networks, even when optimization is unlikely to make a difference. These results suggest that residual connections confer performance advantages beyond optimization, pointing instead to a deeper inductive bias aligned with the structure of natural data. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:33:22 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Mehmeti-Göpel",
"Christian H. X. Ali",
""
],
[
"Wand",
"Michael",
""
]
] |
2506.14387 | William F. Shen | William F. Shen, Xinchi Qiu, Nicola Cancedda, Nicholas D. Lane | Don't Make It Up: Preserving Ignorance Awareness in LLM Fine-Tuning | null | null | null | null | cs.AI | http://creativecommons.org/licenses/by/4.0/ | Existing work on mitigating catastrophic forgetting in large language model (LLM) fine-tuning has primarily focused on preserving specific data or tasks, while critically overlooking the degradation of essential capabilities instilled through safety alignment, particularly the model's ability to faithfully express ignorance. In this work, we show that this capability is significantly degraded during conventional fine-tuning, leading to undesired behaviors such as hallucinations. To address this novel but highly practical problem, we propose SEAT, a simple and effective fine-tuning approach that preserves both fine-tuning performance and the model's inherent ability to acknowledge its ignorance. SEAT integrates two key components: (1) sparse training that constrains activation drift, and (2) a novel entity perturbation method with KL-divergence regularization, designed to counter knowledge entanglement. Experimental results demonstrate that SEAT significantly outperforms baselines in preserving ignorance awareness while retaining fine-tuning performance, offering a more robust solution for LLM fine-tuning. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:33:23 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Shen",
"William F.",
""
],
[
"Qiu",
"Xinchi",
""
],
[
"Cancedda",
"Nicola",
""
],
[
"Lane",
"Nicholas D.",
""
]
] |
2506.14388 | Yota Hino | Y. Hino, Y. Ashida, T. Tano, and Y. Koshio | A simulation model investigation of neutron-oxygen inelastic scattering and subsequent nucleus deexcitation based on experimental data | 7 pages, 5 figures | null | null | null | hep-ex nucl-ex | http://creativecommons.org/licenses/by/4.0/ | The hadron interaction model for light nuclei plays an important role in understanding the event topology of neutrino-nucleus interaction in a large-scale neutrino detector. In the Super-Kamiokande experiment, the neutral current quasi-elastic interaction induced by atmospheric neutrinos introduces 68-82% uncertainty in the search for diffuse supernova neutrino background. This uncertainty stems from the mismodeling of the interaction between oxygen and fast neutrons generated by the neutral current quasi-elastic interaction, which is a crucial factor in the analysis. Therefore, a simulation study was performed, in which the newly released neutron-oxygen experimental data from the E487 and E525 experiments were compared. The result of comparing combinations of intranuclear cascade and deexcitation models is presented to choose a model that provides a superior prediction based on the experimental data. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:35:48 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Hino",
"Y.",
""
],
[
"Ashida",
"Y.",
""
],
[
"Tano",
"T.",
""
],
[
"Koshio",
"Y.",
""
]
] |
2506.14389 | Michael S{\o}rensen | Michael S{\o}rensen | Ole E. Barndorff-Nielsen: Sand, Wind and Inference | null | null | null | null | stat.OT math.ST stat.TH | http://creativecommons.org/licenses/by-nc-sa/4.0/ | This paper reviews Ole Eiler Barndorff-Nielsen's research in the first decades of his career. The focus is on topics that he kept returning to throughout his scientific life, and on papers that he built on in later important contributions. First his early contributions to the foundations of statistical inference are reviewed with focus on conditional inference and exponential families, two topics in which he had a lifelong interest. The second half of the paper reviews his research on wind blown sand and hyperbolic distributions and processes, including his early contributions to modelling of turbulent wind fields. This research laid the foundations for his later work on financial econometrics and ambit processes. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:37:50 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Sørensen",
"Michael",
""
]
] |
2506.14390 | Conrad Orglmeister | Conrad Orglmeister, Erik Bochinski, Volker Eiselein, Elvira Fleig | Enclosing Prototypical Variational Autoencoder for Explainable Out-of-Distribution Detection | This preprint has not undergone peer review or any post-submission improvements or corrections. The Version of Record of this contribution is published in Computer Safety, Reliability and Security - SAFECOMP 2024 Workshops - DECSoS, SASSUR, TOASTS, and WAISE, and is available online at https://doi.org/10.1007/978-3-031-68738-9_29 | null | 10.1007/978-3-031-68738-9_29 | null | cs.LG cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Understanding the decision-making and trusting the reliability of Deep Machine Learning Models is crucial for adopting such methods to safety-relevant applications. We extend self-explainable Prototypical Variational models with autoencoder-based out-of-distribution (OOD) detection: A Variational Autoencoder is applied to learn a meaningful latent space which can be used for distance-based classification, likelihood estimation for OOD detection, and reconstruction. The In-Distribution (ID) region is defined by a Gaussian mixture distribution with learned prototypes representing the center of each mode. Furthermore, a novel restriction loss is introduced that promotes a compact ID region in the latent space without collapsing it into single points. The reconstructive capabilities of the Autoencoder ensure the explainability of the prototypes and the ID region of the classifier, further aiding the discrimination of OOD samples. Extensive evaluations on common OOD detection benchmarks as well as a large-scale dataset from a real-world railway application demonstrate the usefulness of the approach, outperforming previous methods. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:38:29 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Orglmeister",
"Conrad",
""
],
[
"Bochinski",
"Erik",
""
],
[
"Eiselein",
"Volker",
""
],
[
"Fleig",
"Elvira",
""
]
] |
2506.14391 | Yaqiao Zhu | Yaqiao Zhu, Hongkai Wen, Geyong Min, Man Luo | HiLight: A Hierarchical Reinforcement Learning Framework with Global Adversarial Guidance for Large-Scale Traffic Signal Control | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Efficient traffic signal control (TSC) is essential for mitigating urban congestion, yet existing reinforcement learning (RL) methods face challenges in scaling to large networks while maintaining global coordination. Centralized RL suffers from scalability issues, while decentralized approaches often lack unified objectives, resulting in limited network-level efficiency. In this paper, we propose HiLight, a hierarchical reinforcement learning framework with global adversarial guidance for large-scale TSC. HiLight consists of a high-level Meta-Policy, which partitions the traffic network into subregions and generates sub-goals using a Transformer-LSTM architecture, and a low-level Sub-Policy, which controls individual intersections with global awareness. To improve the alignment between global planning and local execution, we introduce an adversarial training mechanism, where the Meta-Policy generates challenging yet informative sub-goals, and the Sub-Policy learns to surpass these targets, leading to more effective coordination. We evaluate HiLight across both synthetic and real-world benchmarks, and additionally construct a large-scale Manhattan network with diverse traffic conditions, including peak transitions, adverse weather, and holiday surges. Experimental results show that HiLight exhibits significant advantages in large-scale scenarios and remains competitive across standard benchmarks of varying sizes. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:39:42 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Zhu",
"Yaqiao",
""
],
[
"Wen",
"Hongkai",
""
],
[
"Min",
"Geyong",
""
],
[
"Luo",
"Man",
""
]
] |
2506.14392 | Rumen Uluchev | Ivan Gadjev, Parvan Parvanov, Rumen Uluchev | Higher Order Approximation of Continuous Functions by a Modified Meyer-K\"{o}nig and Zeller-Type Operator | 18 pages | null | null | null | math.CA | http://creativecommons.org/licenses/by/4.0/ | A new Goodman-Sharma type modification of the Meyer-K\"{o}nig and Zeller operator for approximation of bounded continuous functions on [0,1) is presented. We estimate the approximation error of the proposed operator and prove direct and strong converse theorems with respect to a related K-functional. The operator is linear but not a positive one. However it benefits a better order of approximation compared to the Goodman-Sharma variant of Meyer-K\"{o}nig and Zeller type operator investigated by Ivanov and Parvanov in 2012. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:39:49 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Gadjev",
"Ivan",
""
],
[
"Parvanov",
"Parvan",
""
],
[
"Uluchev",
"Rumen",
""
]
] |
2506.14393 | Kamil Tylinski | Kamil Tylinski, Abylay Satybaldy and Paolo Tasca | Consensus Power Inequality: A Comparative Study of Blockchain Networks | null | null | null | null | cs.CR cs.DC | http://creativecommons.org/licenses/by/4.0/ | The distribution of consensus power is a cornerstone of decentralisation, influencing the security, resilience, and fairness of blockchain networks while ensuring equitable impact among participants. This study provides a rigorous evaluation of consensus power inequality across five prominent blockchain networks - Bitcoin, Ethereum, Cardano, Hedera, and Algorand - using data collected from January 2022 to July 2024. Leveraging established economic metrics, including the Gini coefficient and Theil index, the research quantitatively assesses how power is distributed among blockchain network participants. A robust dataset, capturing network-specific characteristics such as mining pools, staking patterns, and consensus nodes, forms the foundation of the analysis, enabling meaningful comparisons across diverse architectures. Through an in-depth comparative study, the paper identifies key disparities in consensus power distribution. Hedera and Bitcoin demonstrate more balanced power distribution, aligning closely with the principles of decentralisation. Ethereum and Cardano demonstrate moderate levels of inequality. However, contrary to expectations, Ethereum has become more concentrated following its transition to Proof-of-Stake. Meanwhile, Algorand shows a pronounced centralisation of power. Moreover, the findings highlight the structural and operational drivers of inequality, including economic barriers, governance models, and network effects, offering actionable insights for more equitable network design. This study establishes a methodological framework for evaluating blockchain consensus power inequality, emphasising the importance of targeted strategies to ensure fairer power distribution and enhancing the sustainability of decentralised systems. Future research will build on these findings by integrating additional metrics and examining the influence of emerging consensus mechanisms. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:45:41 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Tylinski",
"Kamil",
""
],
[
"Satybaldy",
"Abylay",
""
],
[
"Tasca",
"Paolo",
""
]
] |
2506.14394 | Alexander Christian Mayer | Alexander C. Mayer, Thorsten Naab, Paola Caselli, Alexei V. Ivlev, Tommaso Grassi, Oliver Zier, R\"udiger Pakmor, Stefanie Walch, Volker Springel | Protostellar disks in their natural habitat -- the formation of protostars and their accretion disks in the turbulent and magnetized interstellar medium | 23 pages, 22 figures | null | null | null | astro-ph.GA astro-ph.SR | http://creativecommons.org/licenses/by/4.0/ | We present simulations of the supernova-driven turbulent interstellar medium (ISM) in a simulation domain of volume $(256\,{\rm pc})^3$ within which we resolve the formation of protostellar accretion disks and their stellar cores to spatial scales of $\sim 10^{-4}$ au, using the moving-mesh code {\small AREPO}. We perform simulations with no magnetic fields, ideal magnetohydrodynamics (MHD) and ambipolar diffusion, and compare the resulting first Larson cores and their associated structures, including the accretion disks, their location within the larger-scale structure and the streamers connecting these. We find that disks of sizes $10-100\,{\rm au}$ form early in the simulations without magnetic fields, while there are no disks larger than 10 au with ideal MHD. Ambipolar diffusion causes large disks to form in a subset of cases (two out of six cores), and generally reduces the strength of outflows, which are seen to play a central role. When they are able to carry away significant angular momentum, they prevent the formation of a rotationally supported disk. Magnetic fields strengths grow from $0.1 - 1$ mG in the protostellar core to more than 10 G in the first Larson core in all simulations with ideal MHD. The rotationally supported disks which form can have rotation speeds $> 1$ km s$^{-1}$ even out to further than 100 au from the centre, become gravitationally unstable and form complex spiral substructures with Toomre $Q < 1$. We conclude that the impact of magnetic fields and non-ideal MHD on the formation of protostellar disks is substantial in realistic formation scenarios from the turbulent ISM. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:46:47 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Mayer",
"Alexander C.",
""
],
[
"Naab",
"Thorsten",
""
],
[
"Caselli",
"Paola",
""
],
[
"Ivlev",
"Alexei V.",
""
],
[
"Grassi",
"Tommaso",
""
],
[
"Zier",
"Oliver",
""
],
[
"Pakmor",
"Rüdiger",
""
],
[
"Walch",
"Stefanie",
""
],
[
"Springel",
"Volker",
""
]
] |
2506.14395 | Alexios Liakos | Alexios Liakos | Asteroseismology of four eccentric double-lined spectroscopic eclipsing binaries | 16 pages, 8 figures, 8 tables. Accepted for publication in Astronomy & Astrophysics | null | null | null | astro-ph.SR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Photometric data from the Transiting Exoplanet Survey Satellite (TESS) mission and radial velocities from the Gaia mission and ground-based observations were used to model the light curves and calculate the physical parameters of the eccentric eclipsing systems CH Ind, V577 Oph, CX Phe, and TIC 35481236. The components of these systems have temperatures between 6450 and 7500 K, masses between 1.4 and 1.85 solar masses, and radii between 1.49 and 3.05 solar radii. The residuals of these models were further analyzed using the Fourier method to reveal the pulsational frequencies of their oscillating components. Due to the similarity of the components of each system, the eclipses were used as spatial filters in order to determine which member is the pulsating star. CH Ind was found to pulsate in 46 frequencies; its primary component is a $\gamma$ Dor star and the secondary a $\delta$ Sct star. The primary component of V577 Oph oscillates in both the regimes of $\gamma$ Dor and $\delta$ Sct stars. Moreover, using past timings of minima, an eclipse timing variation analysis was also performed for V577 Oph, resulting in the calculation of the apsidal motion parameters and the existence of a third body around the system. Both components of CX Phe were found to be $\delta$ Sct stars; its primary has three independent frequencies in the range of 14.5-17.4 d$^{-1}$ and its secondary has two main modes of 5.19 and 7.22 d$^{-1}$. The analysis of TIC 35481236 indicates the hybrid $\delta$ Sct-$\gamma$ Dor nature of its secondary component. The physical and pulsational properties of the $\delta$ Sct stars of these systems were compared with those of other $\delta$ Sct stars-members of binaries in evolutionary diagrams. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:50:41 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Liakos",
"Alexios",
""
]
] |
2506.14396 | Jiayi He | Jiayi He, Jiangyan Yi, Jianhua Tao, Siding Zeng, Hao Gu | Manipulated Regions Localization For Partially Deepfake Audio: A Survey | null | null | null | null | cs.SD cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | With the development of audio deepfake techniques, attacks with partially deepfake audio are beginning to rise. Compared to fully deepfake, it is much harder to be identified by the detector due to the partially cryptic manipulation, resulting in higher security risks. Although some studies have been launched, there is no comprehensive review to systematically introduce the current situations and development trends for addressing this issue. Thus, in this survey, we are the first to outline a systematic introduction for partially deepfake audio manipulated region localization tasks, including the fundamentals, branches of existing methods, current limitations and potential trends, providing a revealing insight into this scope. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:51:34 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"He",
"Jiayi",
""
],
[
"Yi",
"Jiangyan",
""
],
[
"Tao",
"Jianhua",
""
],
[
"Zeng",
"Siding",
""
],
[
"Gu",
"Hao",
""
]
] |
2506.14398 | Chia-Hua Wu | Chia-Hua Wu, Wanying Ge, Xin Wang, Junichi Yamagishi, Yu Tsao, Hsin-Min Wang | A Comparative Study on Proactive and Passive Detection of Deepfake Speech | null | null | null | null | cs.SD | http://creativecommons.org/licenses/by/4.0/ | Solutions for defending against deepfake speech fall into two categories: proactive watermarking models and passive conventional deepfake detectors. While both address common threats, their differences in training, optimization, and evaluation prevent a unified protocol for joint evaluation and selecting the best solutions for different cases. This work proposes a framework to evaluate both model types in deepfake speech detection. To ensure fair comparison and minimize discrepancies, all models were trained and tested on common datasets, with performance evaluated using a shared metric. We also analyze their robustness against various adversarial attacks, showing that different models exhibit distinct vulnerabilities to different speech attribute distortions. Our training and evaluation code is available at Github. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:54:08 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Wu",
"Chia-Hua",
""
],
[
"Ge",
"Wanying",
""
],
[
"Wang",
"Xin",
""
],
[
"Yamagishi",
"Junichi",
""
],
[
"Tsao",
"Yu",
""
],
[
"Wang",
"Hsin-Min",
""
]
] |
2506.14399 | Tian Xia | Tian Xia, Fabio De Sousa Ribeiro, Rajat R Rasal, Avinash Kori, Raghav Mehta, Ben Glocker | Decoupled Classifier-Free Guidance for Counterfactual Diffusion Models | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Counterfactual image generation aims to simulate realistic visual outcomes under specific causal interventions. Diffusion models have recently emerged as a powerful tool for this task, combining DDIM inversion with conditional generation via classifier-free guidance (CFG). However, standard CFG applies a single global weight across all conditioning variables, which can lead to poor identity preservation and spurious attribute changes - a phenomenon known as attribute amplification. To address this, we propose Decoupled Classifier-Free Guidance (DCFG), a flexible and model-agnostic framework that introduces group-wise conditioning control. DCFG builds on an attribute-split embedding strategy that disentangles semantic inputs, enabling selective guidance on user-defined attribute groups. For counterfactual generation, we partition attributes into intervened and invariant sets based on a causal graph and apply distinct guidance to each. Experiments on CelebA-HQ, MIMIC-CXR, and EMBED show that DCFG improves intervention fidelity, mitigates unintended changes, and enhances reversibility, enabling more faithful and interpretable counterfactual image generation. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:56:09 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Xia",
"Tian",
""
],
[
"Ribeiro",
"Fabio De Sousa",
""
],
[
"Rasal",
"Rajat R",
""
],
[
"Kori",
"Avinash",
""
],
[
"Mehta",
"Raghav",
""
],
[
"Glocker",
"Ben",
""
]
] |
2506.14400 | Roland Roller | Roland Roller, Michael Hahn, Ajay Madhavan Ravichandran, Bilgin Osmanodja, Florian Oetke, Zeineb Sassi, Aljoscha Burchardt, Klaus Netter, Klemens Budde, Anne Herrmann, Tobias Strapatsas, Peter Dabrock, Sebastian M\"oller | One Size Fits None: Rethinking Fairness in Medical AI | Accepted at the 6th Workshop on Gender Bias in Natural Language Processing at ACL 2025 | null | null | null | cs.LG cs.CY | http://creativecommons.org/licenses/by/4.0/ | Machine learning (ML) models are increasingly used to support clinical decision-making. However, real-world medical datasets are often noisy, incomplete, and imbalanced, leading to performance disparities across patient subgroups. These differences raise fairness concerns, particularly when they reinforce existing disadvantages for marginalized groups. In this work, we analyze several medical prediction tasks and demonstrate how model performance varies with patient characteristics. While ML models may demonstrate good overall performance, we argue that subgroup-level evaluation is essential before integrating them into clinical workflows. By conducting a performance analysis at the subgroup level, differences can be clearly identified-allowing, on the one hand, for performance disparities to be considered in clinical practice, and on the other hand, for these insights to inform the responsible development of more effective models. Thereby, our work contributes to a practical discussion around the subgroup-sensitive development and deployment of medical ML models and the interconnectedness of fairness and transparency. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 10:59:02 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Roller",
"Roland",
""
],
[
"Hahn",
"Michael",
""
],
[
"Ravichandran",
"Ajay Madhavan",
""
],
[
"Osmanodja",
"Bilgin",
""
],
[
"Oetke",
"Florian",
""
],
[
"Sassi",
"Zeineb",
""
],
[
"Burchardt",
"Aljoscha",
""
],
[
"Netter",
"Klaus",
""
],
[
"Budde",
"Klemens",
""
],
[
"Herrmann",
"Anne",
""
],
[
"Strapatsas",
"Tobias",
""
],
[
"Dabrock",
"Peter",
""
],
[
"Möller",
"Sebastian",
""
]
] |
2506.14401 | Sk Mujaffar Hossain | Sk Mujaffer Hossain, Vinila Bedekara, Priyanka Yadavb, Ram Janay Chudhary, Satishchandra Ogale | A Spintronic Battery with Reversible Modulation of Spin Polarization through Li Charge/Discharge: A First Principles Computational Modelling Case Study for an Antiperovskite System | null | null | null | null | cond-mat.mtrl-sci | http://creativecommons.org/licenses/by/4.0/ | A key notion defining the progress of the emergent fields of modern electronics, renewable energy, and smart systems is charge storage, which is primarily embodied in various battery chemistries and systems. In addition to the charge property, the electron also has the spin property, which is exploited in the field of spintronics to access novel magnetically controlled device actions that are not accessible to conventional electronics. An interesting question is whether the two can be fruitfully integrated into a single device concept to expand the horizon of device design and applications. Herein, we present a combined experimental and theoretical study of virgin and lithiated conducting intermetallic anti-perovskite with nominal stoichiometry represented as LixFe3SnC (x = 1, 2, 3, 4) to establish the principle of reversible and concurrent charge and spin polarization storage that can be aptly christened as Iono-Spintronics, representing a notion of a spintronic battery. The experimental results, however, showed that lithiation turns the system into a biphasic state comprised of tin-lithium alloy (due to the high affinity of Sn for Li) along with lithiated Fe3C. The process exhibits multiple cyclability (rechargeability). | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:02:41 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Hossain",
"Sk Mujaffer",
""
],
[
"Bedekara",
"Vinila",
""
],
[
"Yadavb",
"Priyanka",
""
],
[
"Chudhary",
"Ram Janay",
""
],
[
"Ogale",
"Satishchandra",
""
]
] |
2506.14403 | Ding Wang | Ding Wang, Dachuan Feng, Kangcheng Zhou, Yuntian Chen, Shijun Liao, Shiyi Chen | Symbolic Regression-Enhanced Dynamic Wake Meandering: Fast and Physically Consistent Wind-Turbine Wake Modeling | null | null | null | null | physics.flu-dyn | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Accurately modeling wind turbine wakes is essential for optimizing wind farm performance but remains a persistent challenge. While the dynamic wake meandering (DWM) model captures unsteady wake behavior, it suffers from near-wake inaccuracies due to empirical closures. We propose a Symbolic Regression-enhanced DWM (SRDWM) framework that achieves equation-level closure by embedding symbolic expressions for volumetric forcing and boundary terms explicitly into governing equations. These physically consistent expressions are discovered from LES data using symbolic regression guided by a hierarchical, domain-informed decomposition strategy. A revised wake-added turbulence formulation is further introduced to enhance turbulence intensity predictions. Extensive validation across varying inflows shows that SRDWM accurately reproduces both mean wake characteristics and turbulent dynamics, achieving full spatiotemporal resolution with over three orders of magnitude speedup compared to LES. The results highlight symbolic regression as a bridge between data and physics, enabling interpretable and generalizable modeling. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:06:16 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Wang",
"Ding",
""
],
[
"Feng",
"Dachuan",
""
],
[
"Zhou",
"Kangcheng",
""
],
[
"Chen",
"Yuntian",
""
],
[
"Liao",
"Shijun",
""
],
[
"Chen",
"Shiyi",
""
]
] |
2506.14404 | Nikos Spyrou | Nikos Spyrou, Athanasios Vlontzos, Paraskevas Pegios, Thomas Melistas, Nefeli Gkouti, Yannis Panagakis, Giorgos Papanastasiou, Sotirios A. Tsaftaris | Causally Steered Diffusion for Automated Video Counterfactual Generation | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Adapting text-to-image (T2I) latent diffusion models for video editing has shown strong visual fidelity and controllability, but challenges remain in maintaining causal relationships in video content. Edits affecting causally dependent attributes risk generating unrealistic or misleading outcomes if these relationships are ignored. In this work, we propose a causally faithful framework for counterfactual video generation, guided by a vision-language model (VLM). Our method is agnostic to the underlying video editing system and does not require access to its internal mechanisms or finetuning. Instead, we guide the generation by optimizing text prompts based on an assumed causal graph, addressing the challenge of latent space control in LDMs. We evaluate our approach using standard video quality metrics and counterfactual-specific criteria, such as causal effectiveness and minimality. Our results demonstrate that causally faithful video counterfactuals can be effectively generated within the learned distribution of LDMs through prompt-based causal steering. With its compatibility with any black-box video editing system, our method holds significant potential for generating realistic "what-if" video scenarios in diverse areas such as healthcare and digital media. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:06:22 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Spyrou",
"Nikos",
""
],
[
"Vlontzos",
"Athanasios",
""
],
[
"Pegios",
"Paraskevas",
""
],
[
"Melistas",
"Thomas",
""
],
[
"Gkouti",
"Nefeli",
""
],
[
"Panagakis",
"Yannis",
""
],
[
"Papanastasiou",
"Giorgos",
""
],
[
"Tsaftaris",
"Sotirios A.",
""
]
] |
2506.14405 | Jarkko Kotaniemi | Jarkko Kotaniemi, Janne Saukkoriipi, Shuai Li, Markku Suomalainen | Data Driven Approach to Input Shaping for Vibration Suppression in a Flexible Robot Arm | 6 pages, 11 figures, robosoft2025 conference | 2025 IEEE 8th International Conference on Soft Robotics (RoboSoft), Lausanne, Switzerland, 2025, pp. 1-6 | 10.1109/RoboSoft63089.2025.11020930 | null | cs.RO | http://creativecommons.org/licenses/by/4.0/ | This paper presents a simple and effective method for setting parameters for an input shaper to suppress the residual vibrations in flexible robot arms using a data-driven approach. The parameters are adaptively tuned in the workspace of the robot by interpolating previously measured data of the robot's residual vibrations. Input shaping is a simple and robust technique to generate vibration-reduced shaped commands by a convolution of an impulse sequence with the desired input command. The generated impulses create waves in the material countering the natural vibrations of the system. The method is demonstrated with a flexible 3D-printed robot arm with multiple different materials, achieving a significant reduction in the residual vibrations. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:07:28 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Kotaniemi",
"Jarkko",
""
],
[
"Saukkoriipi",
"Janne",
""
],
[
"Li",
"Shuai",
""
],
[
"Suomalainen",
"Markku",
""
]
] |
2506.14406 | Koki Yamauchi | Super-Kamiokande Collaboration: K. Yamauchi, K. Abe, S. Abe, Y. Asaoka, M. Harada, Y. Hayato, K. Hiraide, K. Hosokawa, K. Ieki, M. Ikeda, J. Kameda, Y. Kanemura, Y. Kataoka, S. Miki, S. Mine, M. Miura, S. Moriyama, M. Nakahata, S. Nakayama, Y. Noguchi, G. Pronost, K. Sato, H. Sekiya, K. Shimizu, R. Shinoda, M. Shiozawa, Y. Suzuki, A. Takeda, Y. Takemoto, H. Tanaka, T. Yano, Y. Itow, T. Kajita, R. Nishijima, K. Okumura, T. Tashiro, T. Tomiya, X. Wang, P. Fernandez, L. Labarga, D. Samudio, B. Zaldivar, B. W. Pointon, C. Yanagisawa, E. Kearns, J. Mirabito, L. Wan, T. Wester, J. Bian, B. Cortez, N. J. Griskevich, Y. Jiang, M. B. Smy, H. W. Sobel, V. Takhistov, A. Yankelevich, J. Hill, M. C. Jang, S. H. Lee, D. H. Moon, R. G. Park, B. S. Yang, B. Bodur, K. Scholberg, C. W. Walter, A. Beauch\^ene, O. Drapier, A. Ershova, Th. A. Mueller, A. D. Santos, P. Paganini, C. Quach, R. Rogly, T. Nakamura, J. S. Jang, R. P. Litchfield, L. N. Machado, F. J. P. Soler, J. G. Learned, K. Choi, N. Iovine, D. Tiwari, S. Cao, L. H. V. Anthony, D. Martin, N. W. Prouse, M. Scott, Y. Uchida, V. Berardi, N. F. Calabria, M. G. Catanesi, N. Ospina, E. Radicioni, A. Langella, G. De Rosa, G. Collazuol, M. Feltre, M. Mattiazzi, L. Ludovici, M. Gonin, L. P\'eriss\'e, B. Quilain, S. Horiuchi, A. Kawabata, M. Kobayashi, Y. M. Liu, Y. Maekawa, Y. Nishimura, R. Okazaki, R. Akutsu, M. Friend, T. Hasegawa, Y. Hino, T. Ishida, T. Kobayashi, M. Jakkapu, T. Matsubara, T. Nakadaira, K. Nakamura, Y. Oyama, A. Portocarrero Yrey, K. Sakashita, T. Sekiguchi, T. Tsukamoto, N. Bhuiyan, G. T. Burton, F. Di Lodovico, J. Gao, T. Katori, J. Migenda, R. M. Ramsden, S. Zsoldos, H. Ito, T. Sone, A. T. Suzuki, Y. Takagi, Y. Takeuchi, S. Wada, H. Zhong, J. Feng, L. Feng, S. Han, J. Hikida, J. R. Hu, Z. Hu, M. Kawaue, T. Kikawa, T. Nakaya, T. V. Ngoc, R. A. Wendell, K. Yasutome, S. J. Jenkins, N. McCauley, A. Tarrant, M. Fan\`i, M. J. Wilking, Z. Xie, Y. Fukuda, H. Menjo, Y. Yoshioka, J. Lagoda, M. Mandal, J. Zalipska, M. Mori, M. Jia, J. Jiang, W. Shi, K. Hamaguchi, H. Ishino, Y. Koshio, F. Nakanishi, S. Sakai, T. Tada, T. Tano, T. Ishizuka, G. Barr, D. Barrow, L. Cook, S. Samani, D. Wark, A. Holin, F. Nova, S. Jung, J. Y. Yang, J. Yoo, J. E. P. Fannon, L. Kneale, M. Malek, J. M. McElwee, T. Peacock, P. Stowell, M. D. Thiesse, L. F. Thompson, S. T. Wilson, H. Okazawa, S. M. Lakshmi, E. Kwon, M. W. Lee, J. W. Seo, I. Yu, A. K. Ichikawa, K. D. Nakamura, S. Tairafune, A. Eguchi, S. Goto, S. Kodama, Y. Mizuno, T. Muro, K. Nakagiri, Y. Nakajima, N. Taniuchi, E. Watanabe, M. Yokoyama, P. de Perio, S. Fujita, C. Jes\'us-Valls, K. Martens, Ll. Marti, K. M. Tsui, M. R. Vagins, J. Xia, S. Izumiyama, M. Kuze, R. Matsumoto, K. Terada, R. Asaka, M. Ishitsuka, M. Shinoki, M. Sugo, M. Wako, T. Yoshida, Y. Nakano, F. Cormier, R. Gaur, V. Gousy-Leblanc, M. Hartz, A. Konaka, X. Li, B. R. Smithers, S. Chen, Y. Wu, B. D. Xu, A. Q. Zhang, B. Zhang, M. Girgus, P. Govindaraj, M. Posiadala-Zezula, Y. S. Prabhu, S. B. Boyd, R. Edwards, D. Hadley, M. Nicholson, M. O'Flaherty, B. Richards, A. Ali, B. Jamieson, S. Amanai, C. Bronner, D. Horiguchi, A. Minamino, Y. Sasaki, R. Shibayama, and R. Shimamura | Search for neutron decay into an antineutrino and a neutral kaon in 0.401 megaton-years exposure of Super-Kamiokande | 12 pages, 5 figures | null | null | null | hep-ex | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We searched for bound neutron decay via $n\to\bar{\nu}+K^0$ predicted by the Grand Unified Theories in 0.401 Mton$\cdot$years exposure of all pure water phases in the Super-Kamiokande detector. About 4.4 times more data than in the previous search have been analyzed by a new method including a spectrum fit to kaon invariant mass distributions. No significant data excess has been observed in the signal regions. As a result of this analysis, we set a lower limit of $7.8\times10^{32}$ years on the neutron lifetime at a 90% confidence level. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:08:25 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Kamiokande Collaboration",
"",
""
],
[
"Yamauchi",
"K.",
""
],
[
"Abe",
"K.",
""
],
[
"Abe",
"S.",
""
],
[
"Asaoka",
"Y.",
""
],
[
"Harada",
"M.",
""
],
[
"Hayato",
"Y.",
""
],
[
"Hiraide",
"K.",
""
],
[
"Hosokawa",
"K.",
""
],
[
"Ieki",
"K.",
""
],
[
"Ikeda",
"M.",
""
],
[
"Kameda",
"J.",
""
],
[
"Kanemura",
"Y.",
""
],
[
"Kataoka",
"Y.",
""
],
[
"Miki",
"S.",
""
],
[
"Mine",
"S.",
""
],
[
"Miura",
"M.",
""
],
[
"Moriyama",
"S.",
""
],
[
"Nakahata",
"M.",
""
],
[
"Nakayama",
"S.",
""
],
[
"Noguchi",
"Y.",
""
],
[
"Pronost",
"G.",
""
],
[
"Sato",
"K.",
""
],
[
"Sekiya",
"H.",
""
],
[
"Shimizu",
"K.",
""
],
[
"Shinoda",
"R.",
""
],
[
"Shiozawa",
"M.",
""
],
[
"Suzuki",
"Y.",
""
],
[
"Takeda",
"A.",
""
],
[
"Takemoto",
"Y.",
""
],
[
"Tanaka",
"H.",
""
],
[
"Yano",
"T.",
""
],
[
"Itow",
"Y.",
""
],
[
"Kajita",
"T.",
""
],
[
"Nishijima",
"R.",
""
],
[
"Okumura",
"K.",
""
],
[
"Tashiro",
"T.",
""
],
[
"Tomiya",
"T.",
""
],
[
"Wang",
"X.",
""
],
[
"Fernandez",
"P.",
""
],
[
"Labarga",
"L.",
""
],
[
"Samudio",
"D.",
""
],
[
"Zaldivar",
"B.",
""
],
[
"Pointon",
"B. W.",
""
],
[
"Yanagisawa",
"C.",
""
],
[
"Kearns",
"E.",
""
],
[
"Mirabito",
"J.",
""
],
[
"Wan",
"L.",
""
],
[
"Wester",
"T.",
""
],
[
"Bian",
"J.",
""
],
[
"Cortez",
"B.",
""
],
[
"Griskevich",
"N. J.",
""
],
[
"Jiang",
"Y.",
""
],
[
"Smy",
"M. B.",
""
],
[
"Sobel",
"H. W.",
""
],
[
"Takhistov",
"V.",
""
],
[
"Yankelevich",
"A.",
""
],
[
"Hill",
"J.",
""
],
[
"Jang",
"M. C.",
""
],
[
"Lee",
"S. H.",
""
],
[
"Moon",
"D. H.",
""
],
[
"Park",
"R. G.",
""
],
[
"Yang",
"B. S.",
""
],
[
"Bodur",
"B.",
""
],
[
"Scholberg",
"K.",
""
],
[
"Walter",
"C. W.",
""
],
[
"Beauchêne",
"A.",
""
],
[
"Drapier",
"O.",
""
],
[
"Ershova",
"A.",
""
],
[
"Mueller",
"Th. A.",
""
],
[
"Santos",
"A. D.",
""
],
[
"Paganini",
"P.",
""
],
[
"Quach",
"C.",
""
],
[
"Rogly",
"R.",
""
],
[
"Nakamura",
"T.",
""
],
[
"Jang",
"J. S.",
""
],
[
"Litchfield",
"R. P.",
""
],
[
"Machado",
"L. N.",
""
],
[
"Soler",
"F. J. P.",
""
],
[
"Learned",
"J. G.",
""
],
[
"Choi",
"K.",
""
],
[
"Iovine",
"N.",
""
],
[
"Tiwari",
"D.",
""
],
[
"Cao",
"S.",
""
],
[
"Anthony",
"L. H. V.",
""
],
[
"Martin",
"D.",
""
],
[
"Prouse",
"N. W.",
""
],
[
"Scott",
"M.",
""
],
[
"Uchida",
"Y.",
""
],
[
"Berardi",
"V.",
""
],
[
"Calabria",
"N. F.",
""
],
[
"Catanesi",
"M. G.",
""
],
[
"Ospina",
"N.",
""
],
[
"Radicioni",
"E.",
""
],
[
"Langella",
"A.",
""
],
[
"De Rosa",
"G.",
""
],
[
"Collazuol",
"G.",
""
],
[
"Feltre",
"M.",
""
],
[
"Mattiazzi",
"M.",
""
],
[
"Ludovici",
"L.",
""
],
[
"Gonin",
"M.",
""
],
[
"Périssé",
"L.",
""
],
[
"Quilain",
"B.",
""
],
[
"Horiuchi",
"S.",
""
],
[
"Kawabata",
"A.",
""
],
[
"Kobayashi",
"M.",
""
],
[
"Liu",
"Y. M.",
""
],
[
"Maekawa",
"Y.",
""
],
[
"Nishimura",
"Y.",
""
],
[
"Okazaki",
"R.",
""
],
[
"Akutsu",
"R.",
""
],
[
"Friend",
"M.",
""
],
[
"Hasegawa",
"T.",
""
],
[
"Hino",
"Y.",
""
],
[
"Ishida",
"T.",
""
],
[
"Kobayashi",
"T.",
""
],
[
"Jakkapu",
"M.",
""
],
[
"Matsubara",
"T.",
""
],
[
"Nakadaira",
"T.",
""
],
[
"Nakamura",
"K.",
""
],
[
"Oyama",
"Y.",
""
],
[
"Yrey",
"A. Portocarrero",
""
],
[
"Sakashita",
"K.",
""
],
[
"Sekiguchi",
"T.",
""
],
[
"Tsukamoto",
"T.",
""
],
[
"Bhuiyan",
"N.",
""
],
[
"Burton",
"G. T.",
""
],
[
"Di Lodovico",
"F.",
""
],
[
"Gao",
"J.",
""
],
[
"Katori",
"T.",
""
],
[
"Migenda",
"J.",
""
],
[
"Ramsden",
"R. M.",
""
],
[
"Zsoldos",
"S.",
""
],
[
"Ito",
"H.",
""
],
[
"Sone",
"T.",
""
],
[
"Suzuki",
"A. T.",
""
],
[
"Takagi",
"Y.",
""
],
[
"Takeuchi",
"Y.",
""
],
[
"Wada",
"S.",
""
],
[
"Zhong",
"H.",
""
],
[
"Feng",
"J.",
""
],
[
"Feng",
"L.",
""
],
[
"Han",
"S.",
""
],
[
"Hikida",
"J.",
""
],
[
"Hu",
"J. R.",
""
],
[
"Hu",
"Z.",
""
],
[
"Kawaue",
"M.",
""
],
[
"Kikawa",
"T.",
""
],
[
"Nakaya",
"T.",
""
],
[
"Ngoc",
"T. V.",
""
],
[
"Wendell",
"R. A.",
""
],
[
"Yasutome",
"K.",
""
],
[
"Jenkins",
"S. J.",
""
],
[
"McCauley",
"N.",
""
],
[
"Tarrant",
"A.",
""
],
[
"Fanì",
"M.",
""
],
[
"Wilking",
"M. J.",
""
],
[
"Xie",
"Z.",
""
],
[
"Fukuda",
"Y.",
""
],
[
"Menjo",
"H.",
""
],
[
"Yoshioka",
"Y.",
""
],
[
"Lagoda",
"J.",
""
],
[
"Mandal",
"M.",
""
],
[
"Zalipska",
"J.",
""
],
[
"Mori",
"M.",
""
],
[
"Jia",
"M.",
""
],
[
"Jiang",
"J.",
""
],
[
"Shi",
"W.",
""
],
[
"Hamaguchi",
"K.",
""
],
[
"Ishino",
"H.",
""
],
[
"Koshio",
"Y.",
""
],
[
"Nakanishi",
"F.",
""
],
[
"Sakai",
"S.",
""
],
[
"Tada",
"T.",
""
],
[
"Tano",
"T.",
""
],
[
"Ishizuka",
"T.",
""
],
[
"Barr",
"G.",
""
],
[
"Barrow",
"D.",
""
],
[
"Cook",
"L.",
""
],
[
"Samani",
"S.",
""
],
[
"Wark",
"D.",
""
],
[
"Holin",
"A.",
""
],
[
"Nova",
"F.",
""
],
[
"Jung",
"S.",
""
],
[
"Yang",
"J. Y.",
""
],
[
"Yoo",
"J.",
""
],
[
"Fannon",
"J. E. P.",
""
],
[
"Kneale",
"L.",
""
],
[
"Malek",
"M.",
""
],
[
"McElwee",
"J. M.",
""
],
[
"Peacock",
"T.",
""
],
[
"Stowell",
"P.",
""
],
[
"Thiesse",
"M. D.",
""
],
[
"Thompson",
"L. F.",
""
],
[
"Wilson",
"S. T.",
""
],
[
"Okazawa",
"H.",
""
],
[
"Lakshmi",
"S. M.",
""
],
[
"Kwon",
"E.",
""
],
[
"Lee",
"M. W.",
""
],
[
"Seo",
"J. W.",
""
],
[
"Yu",
"I.",
""
],
[
"Ichikawa",
"A. K.",
""
],
[
"Nakamura",
"K. D.",
""
],
[
"Tairafune",
"S.",
""
],
[
"Eguchi",
"A.",
""
],
[
"Goto",
"S.",
""
],
[
"Kodama",
"S.",
""
],
[
"Mizuno",
"Y.",
""
],
[
"Muro",
"T.",
""
],
[
"Nakagiri",
"K.",
""
],
[
"Nakajima",
"Y.",
""
],
[
"Taniuchi",
"N.",
""
],
[
"Watanabe",
"E.",
""
],
[
"Yokoyama",
"M.",
""
],
[
"de Perio",
"P.",
""
],
[
"Fujita",
"S.",
""
],
[
"Jesús-Valls",
"C.",
""
],
[
"Martens",
"K.",
""
],
[
"Marti",
"Ll.",
""
],
[
"Tsui",
"K. M.",
""
],
[
"Vagins",
"M. R.",
""
],
[
"Xia",
"J.",
""
],
[
"Izumiyama",
"S.",
""
],
[
"Kuze",
"M.",
""
],
[
"Matsumoto",
"R.",
""
],
[
"Terada",
"K.",
""
],
[
"Asaka",
"R.",
""
],
[
"Ishitsuka",
"M.",
""
],
[
"Shinoki",
"M.",
""
],
[
"Sugo",
"M.",
""
],
[
"Wako",
"M.",
""
],
[
"Yoshida",
"T.",
""
],
[
"Nakano",
"Y.",
""
],
[
"Cormier",
"F.",
""
],
[
"Gaur",
"R.",
""
],
[
"Gousy-Leblanc",
"V.",
""
],
[
"Hartz",
"M.",
""
],
[
"Konaka",
"A.",
""
],
[
"Li",
"X.",
""
],
[
"Smithers",
"B. R.",
""
],
[
"Chen",
"S.",
""
],
[
"Wu",
"Y.",
""
],
[
"Xu",
"B. D.",
""
],
[
"Zhang",
"A. Q.",
""
],
[
"Zhang",
"B.",
""
],
[
"Girgus",
"M.",
""
],
[
"Govindaraj",
"P.",
""
],
[
"Posiadala-Zezula",
"M.",
""
],
[
"Prabhu",
"Y. S.",
""
],
[
"Boyd",
"S. B.",
""
],
[
"Edwards",
"R.",
""
],
[
"Hadley",
"D.",
""
],
[
"Nicholson",
"M.",
""
],
[
"O'Flaherty",
"M.",
""
],
[
"Richards",
"B.",
""
],
[
"Ali",
"A.",
""
],
[
"Jamieson",
"B.",
""
],
[
"Amanai",
"S.",
""
],
[
"Bronner",
"C.",
""
],
[
"Horiguchi",
"D.",
""
],
[
"Minamino",
"A.",
""
],
[
"Sasaki",
"Y.",
""
],
[
"Shibayama",
"R.",
""
],
[
"Shimamura",
"R.",
""
]
] |
2506.14407 | Zeinab Sadat Taghavi | Zeinab Sadat Taghavi, Ali Modarressi, Yunpu Ma, Hinrich Sch\"utze | ImpliRet: Benchmarking the Implicit Fact Retrieval Challenge | null | null | null | null | cs.CL cs.AI | http://creativecommons.org/licenses/by/4.0/ | Retrieval systems are central to many NLP pipelines, but often rely on surface-level cues such as keyword overlap and lexical semantic similarity. To evaluate retrieval beyond these shallow signals, recent benchmarks introduce reasoning-heavy queries; however, they primarily shift the burden to query-side processing techniques -- like prompting or multi-hop retrieval -- that can help resolve complexity. In contrast, we present ImpliRet, a benchmark that shifts the reasoning challenge to document-side processing: The queries are simple, but relevance depends on facts stated implicitly in documents through temporal (e.g., resolving "two days ago"), arithmetic, and world knowledge relationships. We evaluate a range of sparse and dense retrievers, all of which struggle in this setting: the best nDCG@10 is only 15.07%. We also test whether long-context models can overcome this limitation. But even with a short context of only ten documents, including the positive document, GPT-4.1 scores only 35.06%, showing that document-side reasoning remains a challenge. Our codes are available at github.com/ZeinabTaghavi/IMPLIRET.Contribution. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:08:29 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Taghavi",
"Zeinab Sadat",
""
],
[
"Modarressi",
"Ali",
""
],
[
"Ma",
"Yunpu",
""
],
[
"Schütze",
"Hinrich",
""
]
] |
2506.14408 | Robert Jack | Himangsu Bhaumik, James E. Hallett, Tanniemola B. Liverpool, Robert L. Jack, C. Patrick Royall | Cyclically sheared colloidal gels: structural change and delayed failure time | 12 pages | null | null | null | cond-mat.soft | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We present experiments and simulations on cyclically sheared colloidal gels, and probe their behaviour on several different length scales. The shearing induces structural changes in the experimental gel, changing particles' neighborhoods and reorganizing the mesoscopic pores. These results are mirrored in computer simulations of a model gel-former, which show how the material evolves down the energy landscape under shearing, for small strains. By systematic variation of simulation parameters, we characterise the structural and mechanical changes that take place under shear, including both yielding and strain-hardening. We simulate creeping flow under constant shear stress, for gels that were previously subject to cyclic shear, showing that strain-hardening also increases gel stability. This response depends on the orientation of the applied shear stress, revealing that the cyclic shear imprints anisotropic structural features into the gel. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:09:33 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Bhaumik",
"Himangsu",
""
],
[
"Hallett",
"James E.",
""
],
[
"Liverpool",
"Tanniemola B.",
""
],
[
"Jack",
"Robert L.",
""
],
[
"Royall",
"C. Patrick",
""
]
] |
2506.14409 | Danilo Monteiro RIbeiro | Rafael C. Lopes, Danilo M. Ribeiro | Defining the Game Producer: A Mapping of Key Characteristics and Differentiators of the Professional Behind Digital Game Production | null | null | null | null | cs.SE | http://creativecommons.org/licenses/by/4.0/ | Introduction: As digital games grow in complexity, the role of the Game Producer becomes increasingly relevant for aligning creative, technical, and business dimensions. Objective: This study aimed to identify and map the main characteristics, skills, and competencies that define the Digital Game Producer profile. Methodology: A qualitative investigation was conducted with 11 semi-structured interviews, analyzed through Grounded Theory to build categories grounded in professional practice. Results: The study produced a structured set of personal characteristics, practical skills, and strategic competencies considered essential for Game Producers. Communication, adaptability, and project management emerged as central elements across the sample. Conclusion: The resulting model offers a foundation for professional training, recruitment strategies, and future research on leadership roles in game development. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:09:47 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Lopes",
"Rafael C.",
""
],
[
"Ribeiro",
"Danilo M.",
""
]
] |
2506.14410 | Tesfa Mengestie | Tesfa Mengestie | Surjective and closed range differentiation operator | null | Rend. Circ. Mat. Palermo, II. Ser, (2024) | 10.1007/s12215-024-01026-2 | null | math.FA | http://creativecommons.org/licenses/by-nc-nd/4.0/ | We identify Fock-type spaces $\mathcal{F}_{(m,p)}$ on which the differentiation operator $D$ has closed range. We prove that $D$ has closed range only if it is surjective, and this happens if and only if
$m=1$. Moreover, since the operator is unbounded on the classical Fock spaces, we consider the modified or the weighted composition--differentiation operator, $D_{(u,\psi,n)} f= u\cdot\big( f^{(n)}\circ \psi\big)$, on these spaces and describe conditions under which the operator admits closed range, surjective, and order bounded structures. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:10:04 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Mengestie",
"Tesfa",
""
]
] |
2506.14411 | John Wikman | John Wikman, Alexandre Proutiere, David Broman | Adaptive Reinforcement Learning for Unobservable Random Delays | null | null | null | null | cs.LG cs.AI cs.RO | http://creativecommons.org/licenses/by/4.0/ | In standard Reinforcement Learning (RL) settings, the interaction between the agent and the environment is typically modeled as a Markov Decision Process (MDP), which assumes that the agent observes the system state instantaneously, selects an action without delay, and executes it immediately. In real-world dynamic environments, such as cyber-physical systems, this assumption often breaks down due to delays in the interaction between the agent and the system. These delays can vary stochastically over time and are typically unobservable, meaning they are unknown when deciding on an action. Existing methods deal with this uncertainty conservatively by assuming a known fixed upper bound on the delay, even if the delay is often much lower. In this work, we introduce the interaction layer, a general framework that enables agents to adaptively and seamlessly handle unobservable and time-varying delays. Specifically, the agent generates a matrix of possible future actions to handle both unpredictable delays and lost action packets sent over networks. Building on this framework, we develop a model-based algorithm, Actor-Critic with Delay Adaptation (ACDA), which dynamically adjusts to delay patterns. Our method significantly outperforms state-of-the-art approaches across a wide range of locomotion benchmark environments. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:11:37 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Wikman",
"John",
""
],
[
"Proutiere",
"Alexandre",
""
],
[
"Broman",
"David",
""
]
] |
2506.14412 | Oleh Astappiev | Tim Cofala, Oleh Astappiev, William Xion, Hailay Teklehaymanot | RAGtifier: Evaluating RAG Generation Approaches of State-of-the-Art RAG Systems for the SIGIR LiveRAG Competition | 4 pages, 5 figures. Report for SIGIR 2025 LiveRAG Challenge | null | null | null | cs.IR cs.AI cs.LG | http://creativecommons.org/licenses/by-sa/4.0/ | Retrieval-Augmented Generation (RAG) enriches Large Language Models (LLMs) by combining their internal, parametric knowledge with external, non-parametric sources, with the goal of improving factual correctness and minimizing hallucinations. The LiveRAG 2025 challenge explores RAG solutions to maximize accuracy on DataMorgana's QA pairs, which are composed of single-hop and multi-hop questions. The challenge provides access to sparse OpenSearch and dense Pinecone indices of the Fineweb 10BT dataset. It restricts model use to LLMs with up to 10B parameters and final answer generation with Falcon-3-10B. A judge-LLM assesses the submitted answers along with human evaluators. By exploring distinct retriever combinations and RAG solutions under the challenge conditions, our final solution emerged using InstructRAG in combination with a Pinecone retriever and a BGE reranker. Our solution achieved a correctness score of 1.13 and a faithfulness score of 0.55, placing fourth in the SIGIR 2025 LiveRAG Challenge. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:14:22 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Cofala",
"Tim",
""
],
[
"Astappiev",
"Oleh",
""
],
[
"Xion",
"William",
""
],
[
"Teklehaymanot",
"Hailay",
""
]
] |
2506.14413 | Jens Gudmundsson | Jens Gudmundsson and Jens Leth Hougaard | Smart contracts and reaction-function games | null | null | null | null | econ.TH | http://creativecommons.org/licenses/by/4.0/ | Blockchain-based smart contracts offer a new take on credible commitment, where players can commit to actions in reaction to actions of others. Such reaction-function games extend on strategic games with players choosing reaction functions instead of strategies. We formalize a solution concept in terms of fixed points for such games, akin to Nash equilibrium, and prove equilibrium existence. Reaction functions can mimic "trigger" strategies from folk theorems on infinitely repeated games -- but now in a one-shot setting. We introduce a refinement in terms of safe play. We apply our theoretical framework to symmetric investment games, which includes two prominent classes of games, namely weakest-link and public-good games. In both cases, we identify a safe and optimal reaction function. In this way, our findings highlight how blockchain-based commitment can overcome trust and free-riding barriers. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:14:57 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Gudmundsson",
"Jens",
""
],
[
"Hougaard",
"Jens Leth",
""
]
] |
2506.14414 | Dawar Khan | Sabahat Israr, Dawar Khan, Zhanglin Cheng, Mukhtaj Khan, Kiyoshi Kiyokawa | GHAR: GeoPose-based Handheld Augmented Reality for Architectural Positioning, Manipulation and Visual Exploration | null | null | null | null | cs.GR | http://creativecommons.org/licenses/by-sa/4.0/ | Handheld Augmented Reality (HAR) is revolutionizing the civil infrastructure application domain. The current trend in HAR relies on marker tracking technology. However, marker-based systems have several limitations, such as difficulty in use and installation, sensitivity to light, and marker design. In this paper, we propose a markerless HAR framework with GeoPose-based tracking. We use different gestures for manipulation and achieve 7 DOF (3 DOF each for translation and rotation, and 1 DOF for scaling). The proposed framework, called GHAR, is implemented for architectural building models. It augments virtual CAD models of buildings on the ground, enabling users to manipulate and visualize an architectural model before actual construction. The system offers a quick view of the building infrastructure, playing a vital role in requirement analysis and planning in construction technology. We evaluated the usability, manipulability, and comprehensibility of the proposed system using a standard user study with the System Usability Scale (SUS) and Handheld Augmented Reality User Study (HARUS). We compared our GeoPose-based markerless HAR framework with a marker-based HAR framework, finding significant improvement in the aforementioned three parameters with the markerless framework. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:17:26 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Israr",
"Sabahat",
""
],
[
"Khan",
"Dawar",
""
],
[
"Cheng",
"Zhanglin",
""
],
[
"Khan",
"Mukhtaj",
""
],
[
"Kiyokawa",
"Kiyoshi",
""
]
] |
2506.14415 | Kamel Khelifa-Kerfa | K. Khelifa-Kerfa and M. Benghanem | Hemisphere mass up to four-loops with generalised $k_t$ algorithms | 12 pages, 4 figures | null | null | null | hep-ph hep-th | http://creativecommons.org/licenses/by/4.0/ | We compute the fixed-order distribution of the non-global hemisphere mass observable in $e^+ e^-$ annihilation up to four loops for various sequential recombination jet algorithms. In particular, we focus on the $k_t$ and Cambridge/Aachen algorithms. Using eikonal theory and strong-energy ordering of the final-state partons, we determine the complete structure of both abelian (clustering) and non-abelian non-global logarithms through four loops in perturbation theory. We compare the resulting resummed expressions for both jet algorithms with the standard Sudakov form factor and demonstrate that neglecting these logarithms leads to unreliable phenomenological predictions for the observable's distribution. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:18:25 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Khelifa-Kerfa",
"K.",
""
],
[
"Benghanem",
"M.",
""
]
] |
2506.14416 | Dana Kamp | Dana Kamp and Karel Proesmans | Thermodynamic control of non-equilibrium systems | 6 pages, 5 figures, 9 page appendix | null | null | null | cond-mat.stat-mech | http://creativecommons.org/licenses/by/4.0/ | We study the thermodynamic cost associated with driving systems between different non-equilibrium steady states. In particular, we combine a linear-response framework for non-equilibrium Markov systems with Lagrangian techniques to minimize the dissipation associated with driving processes. We then apply our framework to a simple toy model. Our results show several remarkable properties for the optimal protocol, such as diverging parameters and finite entropy production in the slow-driving limit. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:26:05 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Kamp",
"Dana",
""
],
[
"Proesmans",
"Karel",
""
]
] |
2506.14417 | Adelle Goodwin Dr | A. J. Goodwin, R. Arcodia, G. Miniutti, J.C.A Miller-Jones, S. van Velzen | The radio properties of quasi-periodic X-ray eruption sources | 15 pages, submitted to PASA, comments welcome | null | null | null | astro-ph.HE | http://creativecommons.org/licenses/by/4.0/ | Quasi-periodic X-ray eruptions (QPEs) are a new class of repeating nuclear transient in which repeating X-ray flares are observed coming from the nuclei of generally low mass galaxies. Here we present a comprehensive summary of the radio properties of 12 bona-fide quasi-periodic eruption sources, including a mix of known tidal disruption events (TDEs) and AGN-like hosts. We include a combination of new dedicated radio observations and archival/previously published radio observations to compile a catalogue of radio observations of each source in the sample. We examine the overall radio properties of the sample and compare to the radio properties of known TDEs, given the apparent link between QPEs and TDEs. Overall we find compact, weak radio sources associated with 5/12 of the QPE sources and no signatures of strong AGN activity via a luminous radio jet. We find no radio variability or correlation between radio emission and the X-ray QPE properties, implying that the mechanism that produces the X-ray flares does not generate strong radio-emitting outflows. The compactness of the radio sources and lack of correlation between radio luminosity and SMBH mass is very unusual for AGN, but the radio spectra and luminosities are consistent with outflows produced by a recent TDE (or accretion event), in both the known TDE sources as well as the AGN-like sources in the sample. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:26:37 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Goodwin",
"A. J.",
""
],
[
"Arcodia",
"R.",
""
],
[
"Miniutti",
"G.",
""
],
[
"Miller-Jones",
"J. C. A",
""
],
[
"van Velzen",
"S.",
""
]
] |
2506.14418 | Yanbiao Ma | Jiayi Chen, Yanbiao Ma, Andi Zhang, Weidong Tang, Wei Dai, Bowei Liu | Compositional Attribute Imbalance in Vision Datasets | null | null | null | null | cs.CV cs.AI | http://creativecommons.org/licenses/by/4.0/ | Visual attribute imbalance is a common yet underexplored issue in image classification, significantly impacting model performance and generalization. In this work, we first define the first-level and second-level attributes of images and then introduce a CLIP-based framework to construct a visual attribute dictionary, enabling automatic evaluation of image attributes. By systematically analyzing both single-attribute imbalance and compositional attribute imbalance, we reveal how the rarity of attributes affects model performance. To tackle these challenges, we propose adjusting the sampling probability of samples based on the rarity of their compositional attributes. This strategy is further integrated with various data augmentation techniques (such as CutMix, Fmix, and SaliencyMix) to enhance the model's ability to represent rare attributes. Extensive experiments on benchmark datasets demonstrate that our method effectively mitigates attribute imbalance, thereby improving the robustness and fairness of deep neural networks. Our research highlights the importance of modeling visual attribute distributions and provides a scalable solution for long-tail image classification tasks. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:28:07 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Chen",
"Jiayi",
""
],
[
"Ma",
"Yanbiao",
""
],
[
"Zhang",
"Andi",
""
],
[
"Tang",
"Weidong",
""
],
[
"Dai",
"Wei",
""
],
[
"Liu",
"Bowei",
""
]
] |
2506.14419 | Cheng Yeaw Ku | Cheng Yeaw Ku and Leyou Xu | Proof of a conjecture on eigenvalues of transposition graph | null | null | null | null | math.CO | http://creativecommons.org/licenses/by/4.0/ | The transposition graph $Cay(S_n,T_n)$ is the Cayley graph on the symmetric group $S_n$ generated by the set $T_n$ of all transpositions. In this paper, we show that each integer in the interval $\left[-{\lfloor(2n+1)/3 \rfloor\choose 2}, {\lfloor(2n+1)/3 \rfloor\choose 2}\right]$ is an eigenvalue of $Cay(S_n,T_n)$. This proves a recent conjecture by Kravchuk \cite{Kravchuk}. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:28:11 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Ku",
"Cheng Yeaw",
""
],
[
"Xu",
"Leyou",
""
]
] |
2506.14420 | Jiakun Zheng | Ting Xiao, Jiakun Zheng, Rushuai Yang, Kang Xu, Qiaosheng Zhang, Peng Liu, Chenjia Bai | Unsupervised Skill Discovery through Skill Regions Differentiation | null | null | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Unsupervised Reinforcement Learning (RL) aims to discover diverse behaviors that can accelerate the learning of downstream tasks. Previous methods typically focus on entropy-based exploration or empowerment-driven skill learning. However, entropy-based exploration struggles in large-scale state spaces (e.g., images), and empowerment-based methods with Mutual Information (MI) estimations have limitations in state exploration. To address these challenges, we propose a novel skill discovery objective that maximizes the deviation of the state density of one skill from the explored regions of other skills, encouraging inter-skill state diversity similar to the initial MI objective. For state-density estimation, we construct a novel conditional autoencoder with soft modularization for different skill policies in high-dimensional space. Meanwhile, to incentivize intra-skill exploration, we formulate an intrinsic reward based on the learned autoencoder that resembles count-based exploration in a compact latent space. Through extensive experiments in challenging state and image-based tasks, we find our method learns meaningful skills and achieves superior performance in various downstream tasks. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:30:04 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Xiao",
"Ting",
""
],
[
"Zheng",
"Jiakun",
""
],
[
"Yang",
"Rushuai",
""
],
[
"Xu",
"Kang",
""
],
[
"Zhang",
"Qiaosheng",
""
],
[
"Liu",
"Peng",
""
],
[
"Bai",
"Chenjia",
""
]
] |
2506.14421 | Dimitrios Thilikos | Laure Morelle and Evangelos Protopapas and Dimitrios M. Thilikos and Sebastian Wiederrecht | Excluding Pinched Spheres | null | null | null | null | math.CO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | The pinched sphere is the pseudo-surface $\mathbb{S}^{\circ}_0$ obtained by identifying two distinct points of the sphere. We provide a structural characterization of graphs excluding an $\mathbb{S}^{\circ}_0$-embeddable graph as a minor. Given a graph $G$ and a vertex set $X$, the bidimensionality of $X$ in $G$ is the maximum $k$ such that $G$ contains the $(k\times k)$-grid as an $X$-rooted minor, i.e., there exists a minor model of the $(k \times k)$-grid in~$G$ such that every branchset of this model contains a vertex of $X$. We prove that there is a function~$f$ such that, if a graph $G$ excludes an $\mathbb{S}^{\circ}_0$-embeddable graph $H$ as a minor, $G$ has a tree decomposition where each torso $G_{t}$ contains some set of vertices $X,$ whose bidimensionality in $G_{t}$ is at most $f(k)$ such that $G_{t}$ can be reduced to a graph embeddable in the projective plane by identifying vertices from $X$. This result is optimal in the sense that every graph admitting such a tree decomposition must exclude some $\mathbb{S}^{\circ}_0$-embeddable graph as a minor. An alternative interpretation of this result can be obtained by the fact that edge-apex graphs, i.e., graphs that can be made planar by removing an edge, are graphs embeddable in the pinched sphere. Several consequences and variants of this min-max duality are discussed. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:30:21 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Morelle",
"Laure",
""
],
[
"Protopapas",
"Evangelos",
""
],
[
"Thilikos",
"Dimitrios M.",
""
],
[
"Wiederrecht",
"Sebastian",
""
]
] |
2506.14422 | Akash Chikhalikar | Akash Chikhalikar, Ankit A. Ravankar, Jose Victorio Salazar Luces and Yasuhisa Hirata | Enhancing Object Search in Indoor Spaces via Personalized Object-factored Ontologies | 8 pages, 9 figures. Accepted for publication in 2025 IEEE/RSJ International Conference on Intelligent Robots and Systems | null | null | null | cs.RO | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Personalization is critical for the advancement of service robots. Robots need to develop tailored understandings of the environments they are put in. Moreover, they need to be aware of changes in the environment to facilitate long-term deployment. Long-term understanding as well as personalization is necessary to execute complex tasks like prepare dinner table or tidy my room. A precursor to such tasks is that of Object Search. Consequently, this paper focuses on locating and searching multiple objects in indoor environments. In this paper, we propose two crucial novelties. Firstly, we propose a novel framework that can enable robots to deduce Personalized Ontologies of indoor environments. Our framework consists of a personalization schema that enables the robot to tune its understanding of ontologies. Secondly, we propose an Adaptive Inferencing strategy. We integrate Dynamic Belief Updates into our approach which improves performance in multi-object search tasks. The cumulative effect of personalization and adaptive inferencing is an improved capability in long-term object search. This framework is implemented on top of a multi-layered semantic map. We conduct experiments in real environments and compare our results against various state-of-the-art (SOTA) methods to demonstrate the effectiveness of our approach. Additionally, we show that personalization can act as a catalyst to enhance the performance of SOTAs. Video Link: https://bit.ly/3WHk9i9 | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:35:10 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Chikhalikar",
"Akash",
""
],
[
"Ravankar",
"Ankit A.",
""
],
[
"Luces",
"Jose Victorio Salazar",
""
],
[
"Hirata",
"Yasuhisa",
""
]
] |
2506.14423 | Andrea Kubin | Andrea Kubin | On variational scheme modeling the anisotropic surface diffusion with elasticity in the plane | null | null | null | null | math.AP | http://creativecommons.org/licenses/by/4.0/ | In this paper, we prove the existence of classical solutions for the anisotropic surface diffusion with elasticity in the plane using a minimizing movements scheme, provided that the initial set is sufficiently regular. This scheme is inspired by the one introduced by Cahn-Taylor [15] to modeling the surface diffusion. Moreover, we prove that this scheme converges to the global solution of the equation. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:36:09 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Kubin",
"Andrea",
""
]
] |
2506.14425 | Tomofumi Kitamura | Tomofumi Kitamura and Alex Fukunaga | Is Selection All You Need in Differential Evolution? | 39 pages, 7 figures | null | null | null | cs.NE cs.AI | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Differential Evolution (DE) is a widely used evolutionary algorithm for black-box optimization problems. However, in modern DE implementations, a major challenge lies in the limited population diversity caused by the fixed population size enforced by the generational replacement. Population size is a critical control parameter that significantly affects DE performance. Larger populations inherently contain a more diverse set of individuals, thereby facilitating broader exploration of the search space. Conversely, when the maximum evaluation budgets is constrained, smaller populations focusing on a limited number of promising candidates may be more suitable. Many state-of-the-art DE variants incorporate an archive mechanism, in which a subset of discarded individuals is preserved in an archive during generation replacement and reused in mutation operations. However, maintaining what is essentially a secondary population via an archive introduces additional design considerations, such as policies for insertion, deletion, and appropriate sizing. To address these limitations, we propose a novel DE framework called Unbounded Differential Evolution (UDE), which adds all generated candidates to the population without discarding any individual based on fitness. Unlike conventional DE, which removes inferior individuals during generational replacement, UDE eliminates replacement altogether, along with the associated complexities of archive management and dynamic population sizing. UDE represents a fundamentally new approach to DE, relying solely on selection mechanisms and enabling a more straightforward yet powerful search algorithm. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:41:44 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Kitamura",
"Tomofumi",
""
],
[
"Fukunaga",
"Alex",
""
]
] |
2506.14426 | Matt Luckcuck | Matt Luckcuck and Angelo Ferrando and Fatma Faruq | Varanus: Runtime Verification for CSP | Accepted at Towards Autonomous Robotic Systems (TAROS) 2025 | null | null | null | cs.LO cs.SE | http://creativecommons.org/licenses/by-nc-sa/4.0/ | Autonomous systems are often used in changeable and unknown environments, where traditional verification may not be suitable. Runtime Verification (RV) checks events performed by a system against a formal specification of its intended behaviour, making it highly suitable for ensuring that an autonomous system is obeying its specification at runtime. Communicating Sequential Processes (CSP) is a process algebra usually used in static verification, which captures behaviour as a trace of events, making it useful for RV as well. Further, CSP has more recently been used to specify autonomous and robotic systems. Though CSP is supported by two extant model checkers, so far it has no RV tool. This paper presents Varanus, an RV tool that monitors a system against an oracle built from a CSP specification. This approach enables the reuse without modifications of a specification that was built, e.g during the system's design. We describe the tool, apply it to a simulated autonomous robotic rover inspecting a nuclear waste store, empirically comparing its performance to two other RV tools using different languages, and demonstrate how it can detect violations of the specification. Varanus can synthesise a monitor from a CSP process in roughly linear time, with respect to the number of states and transitions in the model; and checks each event in roughly constant time. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:42:14 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Luckcuck",
"Matt",
""
],
[
"Ferrando",
"Angelo",
""
],
[
"Faruq",
"Fatma",
""
]
] |
2506.14427 | Shilong Wu | Shilong Wu, Hang Chen, Jun Du | M3SD: Multi-modal, Multi-scenario and Multi-language Speaker Diarization Dataset | 11 pages, 5 figures | null | null | null | eess.AS cs.MM | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | In the field of speaker diarization, the development of technology is constrained by two problems: insufficient data resources and poor generalization ability of deep learning models. To address these two problems, firstly, we propose an automated method for constructing speaker diarization datasets, which generates more accurate pseudo-labels for massive data through the combination of audio and video. Relying on this method, we have released Multi-modal, Multi-scenario and Multi-language Speaker Diarization (M3SD) datasets. This dataset is derived from real network videos and is highly diverse. In addition, we further propose a scenario-related model fine-tuning strategy. Based on the general model pre-trained using the above dataset, we combine the specific data of the target scenario (e.g., meetings) and achieve targeted optimization by using Adapter and LoRA joint fine-tuning, thus achieving the model's domain adaptation. Our dataset and code have been open-sourced at https://huggingface.co/spaces/OldDragon/m3sd. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:44:20 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Wu",
"Shilong",
""
],
[
"Chen",
"Hang",
""
],
[
"Du",
"Jun",
""
]
] |
2506.14428 | Ruihao Xi | Ruihao Xi, Xuekuan Wang, Yongcheng Li, Shuhua Li, Zichen Wang, Yiwei Wang, Feng Wei, Cairong Zhao | Toward Rich Video Human-Motion2D Generation | null | null | null | null | cs.CV | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Generating realistic and controllable human motions, particularly those involving rich multi-character interactions, remains a significant challenge due to data scarcity and the complexities of modeling inter-personal dynamics. To address these limitations, we first introduce a new large-scale rich video human motion 2D dataset (Motion2D-Video-150K) comprising 150,000 video sequences. Motion2D-Video-150K features a balanced distribution of diverse single-character and, crucially, double-character interactive actions, each paired with detailed textual descriptions. Building upon this dataset, we propose a novel diffusion-based rich video human motion2D generation (RVHM2D) model. RVHM2D incorporates an enhanced textual conditioning mechanism utilizing either dual text encoders (CLIP-L/B) or T5-XXL with both global and local features. We devise a two-stage training strategy: the model is first trained with a standard diffusion objective, and then fine-tuned using reinforcement learning with an FID-based reward to further enhance motion realism and text alignment. Extensive experiments demonstrate that RVHM2D achieves leading performance on the Motion2D-Video-150K benchmark in generating both single and interactive double-character scenarios. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:45:33 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Xi",
"Ruihao",
""
],
[
"Wang",
"Xuekuan",
""
],
[
"Li",
"Yongcheng",
""
],
[
"Li",
"Shuhua",
""
],
[
"Wang",
"Zichen",
""
],
[
"Wang",
"Yiwei",
""
],
[
"Wei",
"Feng",
""
],
[
"Zhao",
"Cairong",
""
]
] |
2506.14429 | Xiaoran Liu | Xiaoran Liu, Zhigeng Liu, Zengfeng Huang, Qipeng Guo, Ziwei He, Xipeng Qiu | LongLLaDA: Unlocking Long Context Capabilities in Diffusion LLMs | 16 pages, 12 figures, work in progress | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Large Language Diffusion Models, or diffusion LLMs, have emerged as a significant focus in NLP research, with substantial effort directed toward understanding their scalability and downstream task performance. However, their long-context capabilities remain unexplored, lacking systematic analysis or methods for context extension. In this work, we present the first systematic investigation comparing the long-context performance of diffusion LLMs and traditional auto-regressive LLMs. We first identify a unique characteristic of diffusion LLMs, unlike auto-regressive LLMs, they maintain remarkably \textbf{\textit{stable perplexity}} during direct context extrapolation. Furthermore, where auto-regressive models fail outright during the Needle-In-A-Haystack task with context exceeding their pretrained length, we discover diffusion LLMs exhibit a distinct \textbf{\textit{local perception}} phenomenon, enabling successful retrieval from recent context segments. We explain both phenomena through the lens of Rotary Position Embedding (RoPE) scaling theory. Building on these observations, we propose LongLLaDA, a training-free method that integrates LLaDA with the NTK-based RoPE extrapolation. Our results validate that established extrapolation scaling laws remain effective for extending the context windows of diffusion LLMs. Furthermore, we identify long-context tasks where diffusion LLMs outperform auto-regressive LLMs and others where they fall short. Consequently, this study establishes the first context extrapolation method for diffusion LLMs while providing essential theoretical insights and empirical benchmarks critical for advancing future research on long-context diffusion LLMs. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:45:37 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Liu",
"Xiaoran",
""
],
[
"Liu",
"Zhigeng",
""
],
[
"Huang",
"Zengfeng",
""
],
[
"Guo",
"Qipeng",
""
],
[
"He",
"Ziwei",
""
],
[
"Qiu",
"Xipeng",
""
]
] |
2506.14430 | Eric Jeangirard | Eric Jeangirard | Works-magnet: Accelerating Metadata Curation for Open Science | null | null | null | null | cs.DL | http://creativecommons.org/licenses/by/4.0/ | The transition to Open Science necessitates robust and reliable metadata. While national initiatives, such as the French Open Science Monitor, aim to track this evolution using open data, reliance on proprietary databases persists in many places. Open platforms like OpenAlex still require significant human intervention for data accuracy. This paper introduces Works-magnet, a project by the French Ministry of Higher Education and Research (MESR) Data Science & Engineering Team. Works-magnet is designed to accelerate the curation of bibliographic and research data metadata, particularly affiliations, by making automated AI calculations visible and correctable. It addresses challenges related to metadata heterogeneity, complex processing chains, and the need for human curation in a diverse research landscape. The paper details Works-magnet's concepts, and the observed limitations, while outlining future directions for enhancing open metadata quality and reusability. The works-magnet app is open source on github https://github.com/dataesr/works-magnet | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:46:32 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Jeangirard",
"Eric",
""
]
] |
2506.14431 | Tiantian Zhao | Yong Jiao, Sijie Luo, Tiantian Zhao, Dejian Zhou | Almost uniform convergence for noncommutative Vilenkin-Fourier series | 35 pages | null | null | null | math.FA | http://creativecommons.org/licenses/by/4.0/ | In the present paper, we study almost uniform convergence for noncommutative Vilenkin-Fourier series. Precisely, we establish several noncommutative (asymmetric) maximal inequalities for the Ces\`{a}ro means of the noncommutative Vilenkin-Fourier series, which in turn give the corresponding almost uniform convergence. The primary strategy in our proof is to explore a noncommutative generalization of Sunouchi square function operator, and the very recent advance of the noncommutative Calder\'{o}n-Zygmund decomposition. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:46:40 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Jiao",
"Yong",
""
],
[
"Luo",
"Sijie",
""
],
[
"Zhao",
"Tiantian",
""
],
[
"Zhou",
"Dejian",
""
]
] |
2506.14432 | Asbj{\o}rn Munk | Asbj{\o}rn Munk, Stefano Cerri, Jakob Ambsdorf, Julia Machnio, Sebastian N{\o}rgaard Llambias, Vardan Nersesjan, Christian Hedeager Krag, Peirong Liu, Pablo Rocamora Garc\'ia, Mostafa Mehdipour Ghazi, Mikael Boesen, Michael Eriksen Benros, Juan Eugenio Iglesias and Mads Nielsen | A large-scale heterogeneous 3D magnetic resonance brain imaging dataset for self-supervised learning | null | null | null | null | eess.IV cs.CV | http://creativecommons.org/licenses/by/4.0/ | We present FOMO60K, a large-scale, heterogeneous dataset of 60,529 brain Magnetic Resonance Imaging (MRI) scans from 13,900 sessions and 11,187 subjects, aggregated from 16 publicly available sources. The dataset includes both clinical- and research-grade images, multiple MRI sequences, and a wide range of anatomical and pathological variability, including scans with large brain anomalies. Minimal preprocessing was applied to preserve the original image characteristics while reducing barriers to entry for new users. Accompanying code for self-supervised pretraining and finetuning is provided. FOMO60K is intended to support the development and benchmarking of self-supervised learning methods in medical imaging at scale. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:48:05 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Munk",
"Asbjørn",
""
],
[
"Cerri",
"Stefano",
""
],
[
"Ambsdorf",
"Jakob",
""
],
[
"Machnio",
"Julia",
""
],
[
"Llambias",
"Sebastian Nørgaard",
""
],
[
"Nersesjan",
"Vardan",
""
],
[
"Krag",
"Christian Hedeager",
""
],
[
"Liu",
"Peirong",
""
],
[
"García",
"Pablo Rocamora",
""
],
[
"Ghazi",
"Mostafa Mehdipour",
""
],
[
"Boesen",
"Mikael",
""
],
[
"Benros",
"Michael Eriksen",
""
],
[
"Iglesias",
"Juan Eugenio",
""
],
[
"Nielsen",
"Mads",
""
]
] |
2506.14433 | Dong Li Dr. | Dong Li | Detecting fast-variation pulsations in solar hard X-ray and radio emissions | 5 pages, 5 figures, accepted for publication in MNRAS Letters | null | 10.1093/mnrasl/slaf066 | null | astro-ph.SR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Quasi-periodic pulsations (QPPs) at sub-second periods are frequently detected in the time series of X-rays during stellar flares. However, such rapid pulsations are rarely reported in the hard X-ray (HXR) emission of the small solar flare. We explored the QPP patterns with fast-time variations in HXR and radio emissions produced in a small solar flare on 2025 January 19. By applying the Fast Fourier Transform, the fast-variation pulsations at a quasi-period of about 1 s are identified in the HXR channel of 20-80 keV, which were simultaneously measured by the Hard X-ray Imager and the Konus-Wind. The rapid pulsations with a same quasi-period were also detected in the radio emission at a lower frequency range of about 40-100 MHz. The restructured HXR images show that the QPP patterns mainly locate in footpoint areas that connect by hot plasma loops, and they appear in the flare impulsive phase. Our observations suggest that the fast-variation pulsations could be associated with nonthermal electrons that are periodically accelerated by the intermittent magnetic reconnection, and the 1-s period may be modulated by the coalescence instability between current-carrying loops and magnetic islands. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:49:36 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Li",
"Dong",
""
]
] |
2506.14434 | Bidisha Sharma | Bidisha Sharma, Karthik Pandia Durai, Shankar Venkatesan, Jeena J Prakash, Shashi Kumar, Malolan Chetlur, Andreas Stolcke | Unifying Streaming and Non-streaming Zipformer-based ASR | Accepted in ACL2025 Industry track | null | null | null | cs.SD cs.AI eess.AS | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | There has been increasing interest in unifying streaming and non-streaming automatic speech recognition (ASR) models to reduce development, training, and deployment costs. We present a unified framework that trains a single end-to-end ASR model for both streaming and non-streaming applications, leveraging future context information. We propose to use dynamic right-context through the chunked attention masking in the training of zipformer-based ASR models. We demonstrate that using right-context is more effective in zipformer models compared to other conformer models due to its multi-scale nature. We analyze the effect of varying the number of right-context frames on accuracy and latency of the streaming ASR models. We use Librispeech and large in-house conversational datasets to train different versions of streaming and non-streaming models and evaluate them in a production grade server-client setup across diverse testsets of different domains. The proposed strategy reduces word error by relative 7.9\% with a small degradation in user-perceived latency. By adding more right-context frames, we are able to achieve streaming performance close to that of non-streaming models. Our approach also allows flexible control of the latency-accuracy tradeoff according to customers requirements. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:52:41 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Sharma",
"Bidisha",
""
],
[
"Durai",
"Karthik Pandia",
""
],
[
"Venkatesan",
"Shankar",
""
],
[
"Prakash",
"Jeena J",
""
],
[
"Kumar",
"Shashi",
""
],
[
"Chetlur",
"Malolan",
""
],
[
"Stolcke",
"Andreas",
""
]
] |
2506.14435 | Hongyu Wang | Hongyu Wang, Jiayu Xu, Ruiping Wang, Yan Feng, Yitao Zhai, Peng Pei, Xunliang Cai, Xilin Chen | MoTE: Mixture of Ternary Experts for Memory-efficient Large Multimodal Models | Work in progress | null | null | null | cs.CV cs.LG | http://creativecommons.org/licenses/by/4.0/ | Large multimodal Mixture-of-Experts (MoEs) effectively scale the model size to boost performance while maintaining fixed active parameters. However, previous works primarily utilized full-precision experts during sparse up-cycling. Despite they show superior performance on end tasks, the large amount of experts introduces higher memory footprint, which poses significant challenges for the deployment on edge devices. In this work, we propose MoTE, a scalable and memory-efficient approach to train Mixture-of-Ternary-Experts models from dense checkpoint. Instead of training fewer high-precision experts, we propose to train more low-precision experts during up-cycling. Specifically, we use the pre-trained FFN as a shared expert and train ternary routed experts with parameters in {-1, 0, 1}. Extensive experiments show that our approach has promising scaling trend along model size. MoTE achieves comparable performance to full-precision baseline MoE-LLaVA while offering lower memory footprint. Furthermore, our approach is compatible with post-training quantization methods and the advantage further amplifies when memory-constraint goes lower. Given the same amount of expert memory footprint of 3.4GB and combined with post-training quantization, MoTE outperforms MoE-LLaVA by a gain of 4.3% average accuracy on end tasks, demonstrating its effectiveness and potential for memory-constrained devices. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:53:49 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Wang",
"Hongyu",
""
],
[
"Xu",
"Jiayu",
""
],
[
"Wang",
"Ruiping",
""
],
[
"Feng",
"Yan",
""
],
[
"Zhai",
"Yitao",
""
],
[
"Pei",
"Peng",
""
],
[
"Cai",
"Xunliang",
""
],
[
"Chen",
"Xilin",
""
]
] |
2506.14436 | Shen Yuan | Shen Yuan, Yin Zheng, Taifeng Wang, Binbin Liu, and Hongteng Xu | MoORE: SVD-based Model MoE-ization for Conflict- and Oblivion-Resistant Multi-Task Adaptation | 23 pages, 6 figures | null | null | null | cs.LG | http://creativecommons.org/licenses/by/4.0/ | Adapting large-scale foundation models in multi-task scenarios often suffers from task conflict and oblivion. To mitigate such issues, we propose a novel ''model MoE-ization'' strategy that leads to a conflict- and oblivion-resistant multi-task adaptation method. Given a weight matrix of a pre-trained model, our method applies SVD to it and introduces a learnable router to adjust its singular values based on tasks and samples. Accordingly, the weight matrix becomes a Mixture of Orthogonal Rank-one Experts (MoORE), in which each expert corresponds to the outer product of a left singular vector and the corresponding right one. We can improve the model capacity by imposing a learnable orthogonal transform on the right singular vectors. Unlike low-rank adaptation (LoRA) and its MoE-driven variants, MoORE guarantees the experts' orthogonality and maintains the column space of the original weight matrix. These two properties make the adapted model resistant to the conflicts among the new tasks and the oblivion of its original tasks, respectively. Experiments on various datasets demonstrate that MoORE outperforms existing multi-task adaptation methods consistently, showing its superiority in terms of conflict- and oblivion-resistance. The code of the experiments is available at https://github.com/DaShenZi721/MoORE. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:55:08 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Yuan",
"Shen",
""
],
[
"Zheng",
"Yin",
""
],
[
"Wang",
"Taifeng",
""
],
[
"Liu",
"Binbin",
""
],
[
"Xu",
"Hongteng",
""
]
] |
2506.14437 | Teng Shi | Weicong Qin, Yi Xu, Weijie Yu, Teng Shi, Chenglei Shen, Ming He, Jianping Fan, Xiao Zhang, Jun Xu | Similarity = Value? Consultation Value Assessment and Alignment for Personalized Search | null | null | null | null | cs.IR | http://creativecommons.org/licenses/by/4.0/ | Personalized search systems in e-commerce platforms increasingly involve user interactions with AI assistants, where users consult about products, usage scenarios, and more. Leveraging consultation to personalize search services is trending. Existing methods typically rely on semantic similarity to align historical consultations with current queries due to the absence of 'value' labels, but we observe that semantic similarity alone often fails to capture the true value of consultation for personalization. To address this, we propose a consultation value assessment framework that evaluates historical consultations from three novel perspectives: (1) Scenario Scope Value, (2) Posterior Action Value, and (3) Time Decay Value. Based on this, we introduce VAPS, a value-aware personalized search model that selectively incorporates high-value consultations through a consultation-user action interaction module and an explicit objective that aligns consultations with user actions. Experiments on both public and commercial datasets show that VAPS consistently outperforms baselines in both retrieval and ranking tasks. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:56:11 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Qin",
"Weicong",
""
],
[
"Xu",
"Yi",
""
],
[
"Yu",
"Weijie",
""
],
[
"Shi",
"Teng",
""
],
[
"Shen",
"Chenglei",
""
],
[
"He",
"Ming",
""
],
[
"Fan",
"Jianping",
""
],
[
"Zhang",
"Xiao",
""
],
[
"Xu",
"Jun",
""
]
] |
2506.14438 | Alexis Molina | Pol Ar\'evalo, Alexis Molina, \'Alvaro Ciudad | sHGCN: Simplified hyperbolic graph convolutional neural networks | null | null | null | null | cs.LG cs.AI | http://creativecommons.org/licenses/by/4.0/ | Hyperbolic geometry has emerged as a powerful tool for modeling complex, structured data, particularly where hierarchical or tree-like relationships are present. By enabling embeddings with lower distortion, hyperbolic neural networks offer promising alternatives to Euclidean-based models for capturing intricate data structures. Despite these advantages, they often face performance challenges, particularly in computational efficiency and tasks requiring high precision. In this work, we address these limitations by simplifying key operations within hyperbolic neural networks, achieving notable improvements in both runtime and performance. Our findings demonstrate that streamlined hyperbolic operations can lead to substantial gains in computational speed and predictive accuracy, making hyperbolic neural networks a more viable choice for a broader range of applications. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:58:07 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Arévalo",
"Pol",
""
],
[
"Molina",
"Alexis",
""
],
[
"Ciudad",
"Álvaro",
""
]
] |
2506.14439 | Rikiya Takehi | Rikiya Takehi, Masahiro Asami, Kosuke Kawakami, Yuta Saito | A General Framework for Off-Policy Learning with Partially-Observed Reward | 10 pages, 5 figures. Published as a conference paper at ICLR 2025 | International Conference on Learning Representations (ICLR) 2025 | null | null | cs.LG | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Off-policy learning (OPL) in contextual bandits aims to learn a decision-making policy that maximizes the target rewards by using only historical interaction data collected under previously developed policies. Unfortunately, when rewards are only partially observed, the effectiveness of OPL degrades severely. Well-known examples of such partial rewards include explicit ratings in content recommendations, conversion signals on e-commerce platforms that are partial due to delay, and the issue of censoring in medical problems. One possible solution to deal with such partial rewards is to use secondary rewards, such as dwelling time, clicks, and medical indicators, which are more densely observed. However, relying solely on such secondary rewards can also lead to poor policy learning since they may not align with the target reward. Thus, this work studies a new and general problem of OPL where the goal is to learn a policy that maximizes the expected target reward by leveraging densely observed secondary rewards as supplemental data. We then propose a new method called Hybrid Policy Optimization for Partially-Observed Reward (HyPeR), which effectively uses the secondary rewards in addition to the partially-observed target reward to achieve effective OPL despite the challenging scenario. We also discuss a case where we aim to optimize not only the expected target reward but also the expected secondary rewards to some extent; counter-intuitively, we will show that leveraging the two objectives is in fact advantageous also for the optimization of only the target reward. Along with statistical analysis of our proposed methods, empirical evaluations on both synthetic and real-world data show that HyPeR outperforms existing methods in various scenarios. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 11:58:11 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Takehi",
"Rikiya",
""
],
[
"Asami",
"Masahiro",
""
],
[
"Kawakami",
"Kosuke",
""
],
[
"Saito",
"Yuta",
""
]
] |
2506.14440 | Torbj\"orn Nordling | David E. Hernandez, Jose Chang, Torbj\"orn E. M. Nordling | Model compression using knowledge distillation with integrated gradients | 49 pages, 12 figures | null | null | null | cs.CV cs.AI cs.LG | http://creativecommons.org/licenses/by/4.0/ | Model compression is critical for deploying deep learning models on resource-constrained devices. We introduce a novel method enhancing knowledge distillation with integrated gradients (IG) as a data augmentation strategy. Our approach overlays IG maps onto input images during training, providing student models with deeper insights into teacher models' decision-making processes. Extensive evaluation on CIFAR-10 demonstrates that our IG-augmented knowledge distillation achieves 92.6% testing accuracy with a 4.1x compression factor-a significant 1.1 percentage point improvement ($p<0.001$) over non-distilled models (91.5%). This compression reduces inference time from 140 ms to 13 ms. Our method precomputes IG maps before training, transforming substantial runtime costs into a one-time preprocessing step. Our comprehensive experiments include: (1) comparisons with attention transfer, revealing complementary benefits when combined with our approach; (2) Monte Carlo simulations confirming statistical robustness; (3) systematic evaluation of compression factor versus accuracy trade-offs across a wide range (2.2x-1122x); and (4) validation on an ImageNet subset aligned with CIFAR-10 classes, demonstrating generalisability beyond the initial dataset. These extensive ablation studies confirm that IG-based knowledge distillation consistently outperforms conventional approaches across varied architectures and compression ratios. Our results establish this framework as a viable compression technique for real-world deployment on edge devices while maintaining competitive accuracy. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 12:00:23 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Hernandez",
"David E.",
""
],
[
"Chang",
"Jose",
""
],
[
"Nordling",
"Torbjörn E. M.",
""
]
] |
2506.14441 | Th\'eo Lebeau | Th\'eo Lebeau, Stefano Ettori, Jenny G. Sorce, Nabila Aghanim and Jade Paste | Gas motion in the ICM of the Virgo cluster replica | Submitted to A&A, 17 pages, 9 figures and 1 table | null | null | null | astro-ph.CO astro-ph.GA | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Within the deep gravitational potential of galaxy clusters lies the intracluster medium (ICM). At the first order, it is considered to be at hydrostatic equilibrium within the potential well. However, evidence is growing that the ICM dynamics is non-negligible, is mostly turbulent in origin, and provides a non-thermal pressure support to the equilibrium. In this work, we intend to characterise the properties of the velocity field in the ICM of a simulated replica of the Virgo cluster. We first study the 3D and projected properties of the ICM velocity field by computing its probability density functions (PDFs) and its statistical moments. We then estimate the non-thermal pressure fraction from an effective turbulent Mach number, including the velocity dispersion. We finally compute the velocity structure function (VSF) from projected maps of the sightline velocity. We first show that the components of the 3D velocity field and the projected quantities along equivalent sightlines are anisotropic and affected by the accretion of gas from filaments. Then, we compare the mean statistical moments of the 3D velocity field to the mean properties of a hundred random projections. We show, in particular, an almost linear relation between the standard deviation estimated from direct simulation outputs and sightline velocity dispersion projections, comparable to the line broadening of X-ray atomic lines. However, this linear relation does not hold between the direct simulation outputs and the standard deviation of the sightline velocity projections, comparable to the line shift of X-ray atomic lines. We find a non-thermal pressure fraction around $6\%$ within $R_{500}$ and $9\%$ within $R_{vir}$ from sightline velocity dispersion, which is in good agreement with direct simulation outputs. Finally, we show that the VSF might probe the Active Galactic Nuclei (AGN) feedback turbulent injection scale. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 12:04:18 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Lebeau",
"Théo",
""
],
[
"Ettori",
"Stefano",
""
],
[
"Sorce",
"Jenny G.",
""
],
[
"Aghanim",
"Nabila",
""
],
[
"Paste",
"Jade",
""
]
] |
2506.14442 | Giorgio Di Russo | Donato Bini, Giorgio Di Russo | Topological Stars and scalar wave equation: Exact resummation of the renormalized angular momentum in the eikonal limit | 16 pages, no figures | null | null | null | gr-qc | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | We show that for a Topological Star the renormalized angular momentum parameter, $\nu$, appearing in the Mano-Suzuki-Takasugi-type or in the quantum-Seiberg-Witten-type approaches of the perturbation equations, has 1) a direct link with the geodesic radial action computed along the null orbits of the background and 2) admits an exact resummation in terms of hypergeometric functions, generalizing previous results valid in the Schwarzschild case, see Ref.[arXiv:2504.07862 [hep-th]]. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 12:05:42 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Bini",
"Donato",
""
],
[
"Di Russo",
"Giorgio",
""
]
] |
2506.14443 | Shiba Behera | S. P. Behera, S. Panda, D. K. Mishra | Study of the beyond standard model interaction using Coherent Elastic Neutrino-Nucleus Scattering process | 13 figures, 12 pages | null | null | null | hep-ph hep-ex nucl-ex nucl-th | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Phenomenological studies are carried out to find the potential reach of a detector to constraints on possible beyond the standard model general neutrino interactions in the neutrino-quark sector, for the proposed Indian Coherent Neutrino-nucleus Scattering Experiment (ICNSE) at the Bhabha Atomic Research Centre in India. Reactors of different core configuration and power are used as sources of electron antineutrinos. The study has been carried out considering a target mass of 10 kg and an exposure of one year. It has been found that the ICNSE detector has a significant capacity to limit the majority of the Non-standard Interaction(NSI) parameter space. It has also been observed that a combination of results from two different types of detectors such as sapphire and high-purity germanium can significantly improve sensitivity by reducing the degeneracies between some pairs of NSI parameters to a smaller region. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 12:08:32 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Behera",
"S. P.",
""
],
[
"Panda",
"S.",
""
],
[
"Mishra",
"D. K.",
""
]
] |
2506.14444 | Esin Gulbahar | Esin G. Gulbahar, Camille M. Diez, Aitor Ibarra, Ivan Valtchanov, Richard Saxton, Ignacio de la Calle Perez, Jose Lopez-Miralles, Alejandro Gonzalez Ganzabal, Peter Kretschmar | SAS in ESA Datalabs: A New Platform for XMM-Newton Analysis | 14 pages, 13 figures | Astronomy and Computing, Volume 53, October 2025, 100969 | 10.1016/j.ascom.2025.100969 | null | astro-ph.IM astro-ph.HE | http://creativecommons.org/licenses/by-nc-sa/4.0/ | XMM-Newton is a cornerstone mission of the European Space Agency (ESA) for X-ray astronomy, providing high-quality X-ray data for astrophysical research since the start of the century. Its Science Analysis System (SAS) has been a reliable data reduction and analysis software, evolving throughout the years to meet changing user needs, while incorporating new methods. This paper presents the XMM-SAS Datalab, a tool within the cloud-based ESA Datalabs platform, designed to enhance the interactivity and collaborative potential of SAS. By integrating SAS with a modern, Python-based JupyterLab interface, it enables shared analysis workspaces, removes the need for local software setup, and provides faster access through containerised environments and preconfigured libraries. Moving SAS to the cloud preserves a consistent software setup while eliminating installation complexities, saving time and effort. A case study of the X-ray binary Vela X-1 demonstrates that the Datalabs platform reliably replicates local SAS outputs, with minimal deviations attributed to calibration file versions. The XMM-SAS Datalab allows straightforward X-ray data analysis with collaborative process, setting the way for future adaptations in e-science platforms and multi-wavelength astronomy, while offering traceability and reproducibility of scientific results. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 12:10:08 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Gulbahar",
"Esin G.",
""
],
[
"Diez",
"Camille M.",
""
],
[
"Ibarra",
"Aitor",
""
],
[
"Valtchanov",
"Ivan",
""
],
[
"Saxton",
"Richard",
""
],
[
"Perez",
"Ignacio de la Calle",
""
],
[
"Lopez-Miralles",
"Jose",
""
],
[
"Ganzabal",
"Alejandro Gonzalez",
""
],
[
"Kretschmar",
"Peter",
""
]
] |
2506.14445 | Ruofan Hu | Ruofan Hu, Yan Xia, Minjie Hong, Jieming Zhu, Bo Chen, Xiaoda Yang, Minghui Fang, Tao Jin | Vela: Scalable Embeddings with Voice Large Language Models for Multimodal Retrieval | Accepted by Interspeech 2025 | null | null | null | cs.IR | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | Multimodal large language models (MLLMs) have seen substantial progress in recent years. However, their ability to represent multimodal information in the acoustic domain remains underexplored. In this work, we introduce Vela, a novel framework designed to adapt MLLMs for the generation of universal multimodal embeddings. By leveraging MLLMs with specially crafted prompts and selected in-context learning examples, Vela effectively bridges the modality gap across various modalities. We then propose a single-modality training approach, where the model is trained exclusively on text pairs. Our experiments show that Vela outperforms traditional CLAP models in standard text-audio retrieval tasks. Furthermore, we introduce new benchmarks that expose CLAP models' limitations in handling long texts and complex retrieval tasks. In contrast, Vela, by harnessing the capabilities of MLLMs, demonstrates robust performance in these scenarios. Our code will soon be available. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 12:10:19 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Hu",
"Ruofan",
""
],
[
"Xia",
"Yan",
""
],
[
"Hong",
"Minjie",
""
],
[
"Zhu",
"Jieming",
""
],
[
"Chen",
"Bo",
""
],
[
"Yang",
"Xiaoda",
""
],
[
"Fang",
"Minghui",
""
],
[
"Jin",
"Tao",
""
]
] |
2506.14446 | Nicolau C. Saldanha | Carlos Gustavo Moreira and Nicolau C. Saldanha | Stability of compact actions and a result on divided differences | 20 pages, 1 figure | null | null | null | math.DS | http://creativecommons.org/licenses/by/4.0/ | We study smooth locally free actions of ${\mathbb R}^n$ on manifolds $M$ of dimension $n+1$. We are interested in compact orbits and in compact actions: actions with all orbits compact. Given a compact orbit in a neighborhood of compact orbits, we give necessary and sufficient conditions for the existence of a $C^k$ perturbation with noncompact orbits in the given neighborhood. We prove that if such a perturbation exists it can be assumed to differ from the original action only in a smaller neighborhood of the initial orbit. As an application, for each $k$, we give examples of compact actions which admit $C^{k-1}$-perturbations with noncompact orbits but such that all $C^k$-perturbations are compact. The main result generalizes for $k > 1$ a previous result for the case $C^1$.
A critical auxiliary result is an estimate on divided differences. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 12:11:16 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Moreira",
"Carlos Gustavo",
""
],
[
"Saldanha",
"Nicolau C.",
""
]
] |
2506.14447 | Mikko Partanen | Mikko Partanen, Jukka Tulkki | Perihelion precession of planetary orbits solved from quantum field theory | null | null | null | null | gr-qc | http://creativecommons.org/licenses/by/4.0/ | We derive the perihelion precession of planetary orbits using quantum field theory extending the Standard Model to include gravity. Modeling the gravitational bound state of an electron via the Dirac equation of unified gravity [Rep. Prog. Phys. 88, 057802 (2025)], and taking the classical planetary state limit, we obtain orbital dynamics exhibiting a precession in agreement with general relativity. This demonstrates that key general relativistic effects in planetary motion can emerge directly from quantum field theory without invoking the geometric framework of general relativity. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 12:13:03 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Partanen",
"Mikko",
""
],
[
"Tulkki",
"Jukka",
""
]
] |
2506.14448 | Jiayin Wang | Jiayin Wang, Zhiquang Guo, Weizhi Ma, Min Zhang | How Far Can LLMs Improve from Experience? Measuring Test-Time Learning Ability in LLMs with Human Comparison | null | null | null | null | cs.CL | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ | As evaluation designs of large language models may shape our trajectory toward artificial general intelligence, comprehensive and forward-looking assessment is essential. Existing benchmarks primarily assess static knowledge, while intelligence also entails the ability to rapidly learn from experience. To this end, we advocate for the evaluation of Test-time Learning, the capacity to improve performance in experience-based, reasoning-intensive tasks during test time. In this work, we propose semantic games as effective testbeds for evaluating test-time learning, due to their resistance to saturation and inherent demand for strategic reasoning. We introduce an objective evaluation framework that compares model performance under both limited and cumulative experience settings, and contains four forms of experience representation. To provide a comparative baseline, we recruit eight human participants to complete the same task. Results show that LLMs exhibit measurable test-time learning capabilities; however, their improvements are less stable under cumulative experience and progress more slowly than those observed in humans. These findings underscore the potential of LLMs as general-purpose learning machines, while also revealing a substantial intellectual gap between models and humans, irrespective of how well LLMs perform on static benchmarks. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 12:13:56 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Wang",
"Jiayin",
""
],
[
"Guo",
"Zhiquang",
""
],
[
"Ma",
"Weizhi",
""
],
[
"Zhang",
"Min",
""
]
] |
2506.14449 | Lucas Kreiss | Lucas Kreiss, Amey Chaware, Maryam Roohian, Sarah Lemire, Oana-Maria Thoma, Birgitta Carl\'e, Maximilian Waldner, Sebastian Sch\"urmann, Oliver Friedrich, Roarke Horstmeyer | Detecting immune cells with label-free two-photon autofluorescence and deep learning | null | null | null | null | cs.LG physics.optics | http://creativecommons.org/licenses/by-nc-nd/4.0/ | Label-free imaging has gained broad interest because of its potential to omit elaborate staining procedures which is especially relevant for in vivo use. Label-free multiphoton microscopy (MPM), for instance, exploits two-photon excitation of natural autofluorescence (AF) from native, metabolic proteins, making it ideal for in vivo endomicroscopy. Deep learning (DL) models have been widely used in other optical imaging technologies to predict specific target annotations and thereby digitally augment the specificity of these label-free images. However, this computational specificity has only rarely been implemented for MPM. In this work, we used a data set of label-free MPM images from a series of different immune cell types (5,075 individual cells for binary classification in mixed samples and 3,424 cells for a multi-class classification task) and trained a convolutional neural network (CNN) to classify cell types based on this label-free AF as input. A low-complexity squeezeNet architecture was able to achieve reliable immune cell classification results (0.89 ROC-AUC, 0.95 PR-AUC, for binary classification in mixed samples; 0.689 F1 score, 0.697 precision, 0.748 recall, and 0.683 MCC for six-class classification in isolated samples). Perturbation tests confirmed that the model is not confused by extracellular environment and that both input AF channels (NADH and FAD) are about equally important to the classification. In the future, such predictive DL models could directly detect specific immune cells in unstained images and thus, computationally improve the specificity of label-free MPM which would have great potential for in vivo endomicroscopy. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 12:14:02 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Kreiss",
"Lucas",
""
],
[
"Chaware",
"Amey",
""
],
[
"Roohian",
"Maryam",
""
],
[
"Lemire",
"Sarah",
""
],
[
"Thoma",
"Oana-Maria",
""
],
[
"Carlé",
"Birgitta",
""
],
[
"Waldner",
"Maximilian",
""
],
[
"Schürmann",
"Sebastian",
""
],
[
"Friedrich",
"Oliver",
""
],
[
"Horstmeyer",
"Roarke",
""
]
] |
2506.14450 | Daniel B\"aumer | Daniel B\"aumer, Rupert Klein and Norbert J. Mauser | A general framework for the asymptotic analysis of moist atmospheric flows | null | null | null | null | math-ph math.MP physics.ao-ph | http://creativecommons.org/licenses/by/4.0/ | We deal with asymptotic analysis for the derivation of partial differential equation models for geophysical flows in the earth's atmosphere with moist process closures, and we study their mathematical properties. Starting with the Navier-Stokes equations for dry air, we put the seminal papers of Klein, Majda et al. in a unified context and then discuss the appropriate extension to moist air. In particular, we deal with the scale-independent distinguished limit for the universal parameters of atmospheric motion for moist air, with the Clausius-Clapeyron relation that links saturation vapor pressure and air temperature, and with the mathematical formulation of phase changes associated with cloud formation and rain production. We conclude with a discussion of the precipitating quasigeostrophic (PQG) models introduced by Smith & Stechmann. Our intent is, on the one hand, to convey the problems arising at the modeling stage to mathematicians; on the other hand, we want to present the relevant mathematical methods and results to meteorologists. | [
{
"version": "v1",
"created": "Tue, 17 Jun 2025 12:14:34 GMT"
}
] | 2025-06-18T00:00:00 | [
[
"Bäumer",
"Daniel",
""
],
[
"Klein",
"Rupert",
""
],
[
"Mauser",
"Norbert J.",
""
]
] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.