{
    "paper_id": "S15-1027",
    "header": {
        "generated_with": "S2ORC 1.0.0",
        "date_generated": "2023-01-19T15:36:59.395348Z"
    },
    "title": "Can Selectional Preferences Help Automatic Semantic Role Labeling?",
    "authors": [
        {
            "first": "Shumin",
            "middle": [],
            "last": "Wu",
            "suffix": "",
            "affiliation": {
                "laboratory": "",
                "institution": "University of Colorado",
                "location": {}
            },
            "email": "shumin@colorado.edu"
        },
        {
            "first": "Martha",
            "middle": [],
            "last": "Palmer",
            "suffix": "",
            "affiliation": {
                "laboratory": "",
                "institution": "University of Colorado",
                "location": {}
            },
            "email": "mpalmer@colorado.edu"
        }
    ],
    "year": "",
    "venue": null,
    "identifiers": {},
    "abstract": "We describe a topic model based approach for selectional preference. Using the topic features generated by an LDA model on the extracted predicate-arguments over the Chinese Gigaword corpus, we show improvement to our state-of-the-art Chinese SRL system by 2.34 F1 points on arguments of nominal predicates, 0.40 F1 point on arguments of verb predicates, and 0.66 F1 point overall. More over, similar gains were achieved on out-ofgenre test data, as well as on English SRL using the same technique.",
    "pdf_parse": {
        "paper_id": "S15-1027",
        "_pdf_hash": "",
        "abstract": [
            {
                "text": "We describe a topic model based approach for selectional preference. Using the topic features generated by an LDA model on the extracted predicate-arguments over the Chinese Gigaword corpus, we show improvement to our state-of-the-art Chinese SRL system by 2.34 F1 points on arguments of nominal predicates, 0.40 F1 point on arguments of verb predicates, and 0.66 F1 point overall. More over, similar gains were achieved on out-ofgenre test data, as well as on English SRL using the same technique.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Abstract",
                "sec_num": null
            }
        ],
        "body_text": [
            {
                "text": "It's long been theorized that selectional preferences (SP)/semantic constraints can improve automatic semantic role labeling (SRL). And while there have been several publications showing positive effects of SP, the evaluations have been dominated by pseudodisambiguation. Zapirain et al. (2013) demonstrated end-to-end SRL improvement on arguments of English verb predicates by using a combination of lexical resources and distributional similarity based SP. However, the margin of improvement is a modest 0.4 F1 point (on WSJ) over a baseline system with performance over 4 F1 points lower than the top system in CoNLL-2005 (Carreras and M\u00e0rquez, 2005) . These results may not be convincing enough to motivate the incorporation of SP when building an SRL system. One reason for the small improvement may be that arguments of a verb predicate are highly constrained by the underlying syntactic parse, and SP features that could disambiguate between role types are often negated by parse errors. With the recent extension of PropBank SRL to nominal and adjective predicates, preposition relationships, light-verb constructions, and abstract meaning representation (Bonial et al., 2014; Banarescu et al., 2013) , it may be time to revisit SP for SRL. We hypothesize that SP will provide a greater benefit to nominal SRL, especially on a language with lower parsing accuracy.",
                "cite_spans": [
                    {
                        "start": 614,
                        "end": 624,
                        "text": "CoNLL-2005",
                        "ref_id": null
                    },
                    {
                        "start": 625,
                        "end": 653,
                        "text": "(Carreras and M\u00e0rquez, 2005)",
                        "ref_id": "BIBREF3"
                    },
                    {
                        "start": 1163,
                        "end": 1184,
                        "text": "(Bonial et al., 2014;",
                        "ref_id": "BIBREF2"
                    },
                    {
                        "start": 1185,
                        "end": 1208,
                        "text": "Banarescu et al., 2013)",
                        "ref_id": "BIBREF0"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "In this paper, we apply SP to Chinese SRL (which has few morphological clues that impacts parsing accuracy) for arguments of both verb and nominal predicates using Chinese Gigaword. Our hypothesis, that SP will provide a greater benefit for nominal predicates than for verbal predicates, is verified by our results. We achieve a 2.34 F1 point improvement to our Chinese SRL system on arguments of nominal predicates, 0.40 F1 point on arguments of verb predicates, and 0.66 F1 point overall.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "Inducing selectional preferences from corpus data was first proposed by Resnik (1997) for sense disambiguation. He generalized seen words using the WordNet (Fellbaum, 1998) hierarchy. Gildea and Jurafsky (2002) applied SP to automatic SRL by clustering extracted verb-direct object pairs, resulting in modest improvements. This syntactic signature based selectional preference technique has also been successfully extended and applied to unsupervised SRL by Lang and Lapata (2011) (using splitmerge role clustering), as well as Titov and Klementiev (2012) (using a distance-dependent Chinese Restaurant Process prior for role clustering). Zapirain et al. (2013) improved the end-to-end perfor-mance of an English PropBank SRL system by 0.4 F1 points using a variety of word similarity measures, from WordNet hierarchy distance to distributional similarity measures. Ritter and Etzioni (2010) reasoned that the set of hidden variables modeled by latent Dirichlet allocation (LDA) naturally represents the semantic structure of a document collection, and the topics generated can be viewed as the latent set of classes that store preferences. The work utilizes LinkLDA, a variant of the standard LDA that models two sets of distributions for each topic simultaneously, with the resulting topics encoding the mutual constraints of a pair of arguments for the same predicate. S\u00e9aghdha and Korhonen (2014) also proposed SP w/ the LDA variants ROOTH-LDA and LEX-LDA.",
                "cite_spans": [
                    {
                        "start": 72,
                        "end": 85,
                        "text": "Resnik (1997)",
                        "ref_id": "BIBREF13"
                    },
                    {
                        "start": 156,
                        "end": 172,
                        "text": "(Fellbaum, 1998)",
                        "ref_id": "BIBREF6"
                    },
                    {
                        "start": 184,
                        "end": 210,
                        "text": "Gildea and Jurafsky (2002)",
                        "ref_id": "BIBREF7"
                    },
                    {
                        "start": 639,
                        "end": 661,
                        "text": "Zapirain et al. (2013)",
                        "ref_id": "BIBREF20"
                    },
                    {
                        "start": 866,
                        "end": 891,
                        "text": "Ritter and Etzioni (2010)",
                        "ref_id": "BIBREF14"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Previous Work on Selectional Preference",
                "sec_num": "2"
            },
            {
                "text": "There has also been work on Chinese selectional preferences, both lexical resource (HowNet) based and corpus based (Jia et al., 2011; Jia et al., 2013) . The authors found the LDA corpus based SP improved over the HowNet based SP on pseudodisambiguation. All of these results encouraged us to also attempt an LDA based approach to SP.",
                "cite_spans": [
                    {
                        "start": 115,
                        "end": 133,
                        "text": "(Jia et al., 2011;",
                        "ref_id": "BIBREF8"
                    },
                    {
                        "start": 134,
                        "end": 151,
                        "text": "Jia et al., 2013)",
                        "ref_id": "BIBREF9"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Previous Work on Selectional Preference",
                "sec_num": "2"
            },
            {
                "text": "3 Selectional Preference for SRL",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Previous Work on Selectional Preference",
                "sec_num": "2"
            },
            {
                "text": "Some of the most discriminative SP models used by Zapirain et al. (2013) relied on distributional similarity computed over dependency relationships (provided by Lin (1998) ). For example, in \"John lent Mary the book.\", we would extract John-nsubj, Mary-iobj, book-dobj for the predicate lend. While this has proven to be of higher quality than pure word co-occurrence based similarity, it may not be optimal for semantic-based processing. With nominal SRL, a large portion of the arguments (around 50% in Chinese PropBank) are not the direct syntactic dependents of the predicate: in figure 1, because of a light verb-like construction, all the arguments of \u6b22\u8fce/welcome are the syntactic dependents of \u8868 \u793a/express. To address this, we directly extract SP of the predicates by running our SRL system over the unannotated corpus. For our example, we would extract John-Arg0, Mary-Arg2, book-Arg1 for lend.",
                "cite_spans": [
                    {
                        "start": 161,
                        "end": 171,
                        "text": "Lin (1998)",
                        "ref_id": "BIBREF11"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "SP Representation",
                "sec_num": "3.1"
            },
            {
                "text": "Our approach to modeling selectional preferences (SP) follows a relatively straightforward application of LDA to a set of predicate-argument instances derived from a corpus. In the standard LDA model, a document d is represented by a bag of words and is drawn from a multi-nominal Dirichlet \u03b8 d over topics. The resulting model is a probability distribution of each word amongst the topics.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "SP with LDA-based Topic Model",
                "sec_num": "3.2"
            },
            {
                "text": "For the SRL application, we treat each extracted argument (represented by the (label, headword) pair) as a \"word\", and the collection of arguments for all instances of a particular predicate as a \"document\". The generated topics would then contain arguments sharing a similar set of predicates. With this definition, we allow different role labels to share the same topic (though it does not encode role constraints quite like LinkLDA, ROOTH-LDA, etc). For prepositional phrases, we used the dependent of the preposition as the head word since the preposition can often be omitted in Chinese.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "SP with LDA-based Topic Model",
                "sec_num": "3.2"
            },
            {
                "text": "Building selectional preferences by means of using the output of an SRL system is unlikely to improve the same SRL system unless one filters out the lower quality labels (in earlier experiments where we performed no filtering, this was indeed the case). We ran SRL on the unannotated corpus using a logistic regression model and filtered out the low probability output. To balance between precision and recall, we set a hard 0.5 probability cutoff and discounted the occurrences of the rest using the label probability.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "SRL Filtering",
                "sec_num": "3.3"
            },
            {
                "text": "Since we can extract higher quality SP from the output of a better performing SRL system, we can iteratively improve our SRL system by re-extracting SP using a retrained (SP enhanced) SRL system. We arrived at diminishing returns after one additional iteration (of training SRL, extracting SP, and retraining SRL w/ new SP).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "SRL Filtering",
                "sec_num": "3.3"
            },
            {
                "text": "Our Chinese SRL system follows the standard (English) approach where the SRL task is posed as a multi-class classification problem requiring the identification of argument candidates for each predicate and their argument types using a set of lexical and syntactic features (predicate word, constituent head, path, syntactic frame, etc). While the top SRL systems from CoNLL-2005 1 and some subsequent systems use multiple parses for structural inference, we instead implement a 2-stage argument label classification system on a single input parse: the argument set found by the first classifier is used as an additional feature for the second classifier (to identify missing or duplicate argument label types).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "SRL Implementation",
                "sec_num": "4"
            },
            {
                "text": "The LDA topic model produces a probability distribution of words (represented here by the (label, headword) pair) over topics. For the SRL task, argument candidates with topic distributions similar to those of the arguments found in the training set are likely to be permissible. Ideally, we would use these distributions directly. Since our SRL system was designed to accept lexical (binary) features only (for training/decoding performance), we pared the distribution down to at most 3 topics for each label type and excluded words that do not have high affinity to a few topics (sum of the probability of the top 3 topics < 50%) to prevent diluting the discriminative power of the topic feature. We used the resulting list of (label, topic id) pairs for each word as the selectional preference feature for each encountered constituent in the Chinese SRL system. During the normal LDA inference stage, using the learned topic model, a predicate instance (\"document\") will be assigned a probability distribution over topics based on its arguments, and each argument will be assigned a specific topic (or topic distribution). This could further constrain an argument's selectional preference within the context of the predicate instance and other arguments. For our system, we experimented with performing inference on the argument label set extracted from the first stage classifier and using the constrained argument topic dis-tribution for the second stage classifier. However, we observed no improvement, likely because there are only a few arguments for each predicate instance.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Selectional Preference",
                "sec_num": "4.1"
            },
            {
                "text": "Our Chinese SRL system is trained on Chinese Tree-Bank 5.1 and Chinese PropBank 1.0. We used the standard: sections 81-885 for training, sections 41-80 for development, and sections 1-40, 900-931 for testing. We generated the training parses (with 10 fold cross-validation) and the test parses using the Berkeley parser 2 (5 split-merge cycles). The parser F1 score on the test sections is 82.73 as measured by ParseEval (Black et al., 1991) .",
                "cite_spans": [
                    {
                        "start": 421,
                        "end": 441,
                        "text": "(Black et al., 1991)",
                        "ref_id": "BIBREF1"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Setup",
                "sec_num": "5.1"
            },
            {
                "text": "We prepared the Chinese Gigaword 3 corpus with the Stanford Chinese Word Segmenter 4 . We performed LDA topic modeling using PLDA+ (Liu et al., 2011) and the recommended \u03b1 = 50/topic cnt, \u03b2 = 0.01 values. We chose 2000 topics (tuned on the SRL performance of the development set rather than any topic based metrics). Table 1 lists some of the found topics (with the most frequent, relatively interesting, and least frequent headword, label pairs) using Chinese Gigaword.",
                "cite_spans": [
                    {
                        "start": 131,
                        "end": 149,
                        "text": "(Liu et al., 2011)",
                        "ref_id": "BIBREF12"
                    }
                ],
                "ref_spans": [
                    {
                        "start": 317,
                        "end": 324,
                        "text": "Table 1",
                        "ref_id": "TABREF1"
                    }
                ],
                "eq_spans": [],
                "section": "Setup",
                "sec_num": "5.1"
            },
            {
                "text": "As table 2 shows, the addition of the SP feature improved nominal SRL by 2.34 F1 points. Verb SRL improved by 0.40 F1 point and overall SRL improved by 0.66 F1 point. These F1 differences were all found to be statistically significant 5 (p \u2264 0.05).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Performance",
                "sec_num": "5.2"
            },
            {
                "text": "We also tested the system on Sinorama magazine and other out-of-genre sections (broadcast conversation, broadcast news, web blog) in Chinese Prop-topic headword:argument label pairs emergency response \u7834\u574f/damage:Arg1 \u963b\u6b62/stop:Arg1 \u5236\u9020/fabricate:Arg1 \u5bfb\u627e/search:Arg1 \u81ea\u6740/suicide:Arg1",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Performance",
                "sec_num": "5.2"
            },
            {
                "text": "... \u706d\u706b/extinguish:Arg1 \u6572\u8bc8/blackmail:Arg1 \u6323\u8131/break free:Arg1 \u4e1c\u5c71\u518d\u8d77/comeback:Arg1 government agency",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Performance",
                "sec_num": "5.2"
            },
            {
                "text": "\u6d77 \u5173/custom:Arg0 \u8054 \u5408 \u4f1a/union:Arg0 \u52a1 \u90e8/work department:Arg0 \u65c5 \u6e38 \u5c40/travel department:Arg0 \u7edf \u8ba1 \u5c40/census:Arg0 ... \u90e8 \u4f1a/ministries:Arg0 \u8fb9 \u68c0 \u7ad9/checkpoint:Arg0 \u8d22\u653f\u5c40/finance bureau:Arg0",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Performance",
                "sec_num": "5.2"
            },
            {
                "text": "law & order \u8b66 \u65b9/police:Arg0 \u5acc \u72af/suspect:Arg1 \u7537 \u5b50/male:Arg1 \u5230 \u6848/court appearance:Arg1 \u516c \u5b89/public safety:Arg0 ... \u5df7/alley:Argm-loc \u5609 \u4e49 \u5e02/Chiayi City:Argm-loc \u54e5 \u4f26 \u6bd4 \u4e9a \u4eba/Columbian:Arg1 path \u9053 \u8def/road:Arg1 \u8def/path:Arg1 \u5927 \u9053/avenue:Arg1 ... \u7ea2 \u5730 \u6bef/red carpet:Arg1 \u94a2 \u4e1d/steel wire:Arg1 \u72ec\u6728\u6865/plank bridge:Arg1 ... \u8ff7\u5bab/maze:Arg1 \u4fa7\u95e8/side entrance:Arg1 \u9669 \u68cb/risky move:Arg1 competition \u6bd4 \u8d5b/competition:Arg1 \u51b3 \u8d5b/final:Arg1 \u8054 \u8d5b/league comp:Arg1 ... \u8003 \u8bd5/exam:Arg1 \u5927 \u9009/election:Arg1 \u4e16 \u4e52 \u8d5b/world pingpong match:Arg1 ...",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Performance",
                "sec_num": "5.2"
            },
            {
                "text": "\u52a0 \u8d5b/playoff:Arg1 \u5206 \u56e2/sub- group:Arg0 moral & ethics",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Performance",
                "sec_num": "5.2"
            },
            {
                "text": "\u7cbe \u795e/spirit:Arg1 \u4f20 \u7edf/tradition:Arg1 \u4f5c \u98ce/style:Arg1 \u6587 \u660e/civil:Arg1 ... \u6821 \u98ce/school spirit:Arg1 \u540c \u821f \u5171 \u6d4e/share hard time:Arg1 ... \u5e78 \u798f \u89c2/happy outlook:Arg1 \u535a \u7231/universal love:Arg1 Bank 3.0. Only Sinorama has nominal SRL annotations. As table 3 shows, even though the absolute performance is much lower, SP improved the precision and recall in all cases, the nominal SRL score on Sinorama by 2.30 F1 points, and verb SRL score by 0.31-0.46 F1 point. Again, these F1 differences were statistically significant.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Performance",
                "sec_num": "5.2"
            },
            {
                "text": "Direct performance comparison with previous Chinese SRL systems is a bit difficult: Xue (2008) , Zhuang and Zong (2010) trained the syntactic parsers with an additional 250K word broadcast news corpus found in Chinese TreeBank 6.0, while Sun (2010) only reported results using gold POS tags but no additional gold parses. However, as table 4 shows, for verb predicates, our system bests Xue's (2008) system by 4-7 F1 points with less parser training data and when tested with (but was not retrained to take full advantage of) gold POS tags besting Sun's (2010) system by 0.53 F1 point. For nominal predicates, our system bests Xue's (2008) system, by 1.9 F1 points on arguments of nominal predicates (since we have an integrated SRL system, the results are obtained by training both verb and nominal predicates, then using only the nominal classifier to classify the nominal predicates).",
                "cite_spans": [
                    {
                        "start": 84,
                        "end": 94,
                        "text": "Xue (2008)",
                        "ref_id": "BIBREF18"
                    },
                    {
                        "start": 97,
                        "end": 119,
                        "text": "Zhuang and Zong (2010)",
                        "ref_id": "BIBREF21"
                    },
                    {
                        "start": 387,
                        "end": 399,
                        "text": "Xue's (2008)",
                        "ref_id": "BIBREF18"
                    },
                    {
                        "start": 627,
                        "end": 639,
                        "text": "Xue's (2008)",
                        "ref_id": "BIBREF18"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Comparison",
                "sec_num": "5.2.1"
            },
            {
                "text": "We applied the same techniques to English SRL using the English Gigaword 7 corpus. We used 800 topics (w/ lemmatized headwords) tuning on the 5) , our SP approach had a smaller (but still statistically significant) absolute F1 gain, with most of the gain coming from core argument type improvements. But with a much higher performing baseline system (one of the highest reported results using a single input parse per sentence), the error reduction rate is comparable.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 142,
                        "end": 144,
                        "text": "5)",
                        "ref_id": "TABREF6"
                    }
                ],
                "eq_spans": [],
                "section": "English SRL",
                "sec_num": "5.2.2"
            },
            {
                "text": "We presented a LDA topic model based selectional preference approach to improving automatic SRL. Using SP extracted from a 63.6M sentence Chinese Gigaword corpus, we were able to improve on the results of an already competitive Chinese SRL system by 2.34 F1 points on nominal predicates, 0.40 F1 point on verb predicates, and 0.66 F1 point on the standard test set. More over, we obtained comparable improvement on out-of-genre data and demonstrated our technique is also applicable to English SRL. Given the margin of improvement on nominal SRL, which is not as well constrained by syntax as verb SRL, there are reasons to speculate the proposed technique could be applicable to other predicate type extensions of PropBank SRL. As our first attempt at automatically deriving Chinese selectional preference, there is a lot of room for future improvement. Notably, these include techniques used for English SP such as computing similarity based on lexical resources (for Chinese -HowNet (Dong et al., 2010) ), distributional similarity, latent word language model (Deschacht and Moens, 2009) , different variants of LDA topic models, as well as taking advantages of argument constraints in parallel corpora to extract higher quality SP.",
                "cite_spans": [
                    {
                        "start": 986,
                        "end": 1005,
                        "text": "(Dong et al., 2010)",
                        "ref_id": "BIBREF5"
                    },
                    {
                        "start": 1063,
                        "end": 1090,
                        "text": "(Deschacht and Moens, 2009)",
                        "ref_id": "BIBREF4"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Conclusion",
                "sec_num": "6"
            },
            {
                "text": "We use CoNLL-2005 instead of CoNLL-2009 for comparison because our SRL system is based on constituent parses.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            },
            {
                "text": "code.google.com/p/berkeleyparser/ 3 LDC2011T13 4 nlp.stanford.edu/software/segmenter.shtml 5 SIGF (www.nlpado.de/%7esebastian/software/sigf.shtml), using stratified approximate randomization test(Yeh, 2000)",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            },
            {
                "text": "Verb results are from SRL systems trained on verbs only.Table 2 results are from SRL systems trained on all predicates. 7 LDC2003T05",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            }
        ],
        "back_matter": [
            {
                "text": "We gratefully acknowledge the support of the National Science Foundation CISE-IISRI-0910992, Richer Representations for Machine Translation, DARPA FA8750-09-C-0179 (via BBN) Machine Reading: Ontology Induction: Semlink+, and DARPA HR0011-11-C-0145 (via LDC) BOLT. This work utilized the Janus supercomputer, which is supported by the National Science Foundation (award number CNS-0821794) and the University of Colorado Boulder. The Janus supercomputer is a joint effort of the University of Colorado Boulder, the University of Colorado Denver and the National Center for Atmospheric Research. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Acknowledgement",
                "sec_num": null
            }
        ],
        "bib_entries": {
            "BIBREF0": {
                "ref_id": "b0",
                "title": "Proceedings of the 7th Linguistic Annotation Workshop and Interoperability with Discourse, chapter Abstract Meaning Representation for Sembanking",
                "authors": [
                    {
                        "first": "Laura",
                        "middle": [],
                        "last": "Banarescu",
                        "suffix": ""
                    },
                    {
                        "first": "Claire",
                        "middle": [],
                        "last": "Bonial",
                        "suffix": ""
                    },
                    {
                        "first": "Shu",
                        "middle": [],
                        "last": "Cai",
                        "suffix": ""
                    },
                    {
                        "first": "Madalina",
                        "middle": [],
                        "last": "Georgescu",
                        "suffix": ""
                    },
                    {
                        "first": "Kira",
                        "middle": [],
                        "last": "Griffitt",
                        "suffix": ""
                    },
                    {
                        "first": "Ulf",
                        "middle": [],
                        "last": "Hermjakob",
                        "suffix": ""
                    },
                    {
                        "first": "Kevin",
                        "middle": [],
                        "last": "Knight",
                        "suffix": ""
                    },
                    {
                        "first": "Philipp",
                        "middle": [],
                        "last": "Koehn",
                        "suffix": ""
                    },
                    {
                        "first": "Martha",
                        "middle": [],
                        "last": "Palmer",
                        "suffix": ""
                    },
                    {
                        "first": "Nathan",
                        "middle": [],
                        "last": "Schneider",
                        "suffix": ""
                    }
                ],
                "year": 2013,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "178--186",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Laura Banarescu, Claire Bonial, Shu Cai, Madalina Georgescu, Kira Griffitt, Ulf Hermjakob, Kevin Knight, Philipp Koehn, Martha Palmer, and Nathan Schneider, 2013. Proceedings of the 7th Linguistic Annotation Workshop and Interoperability with Dis- course, chapter Abstract Meaning Representation for Sembanking, pages 178-186. Association for Com- putational Linguistics.",
                "links": null
            },
            "BIBREF1": {
                "ref_id": "b1",
                "title": "Procedure for quantitatively comparing the syntactic coverage of english grammars",
                "authors": [
                    {
                        "first": "E",
                        "middle": [],
                        "last": "Black",
                        "suffix": ""
                    },
                    {
                        "first": "S",
                        "middle": [],
                        "last": "Abney",
                        "suffix": ""
                    },
                    {
                        "first": "S",
                        "middle": [],
                        "last": "Flickenger",
                        "suffix": ""
                    },
                    {
                        "first": "C",
                        "middle": [],
                        "last": "Gdaniec",
                        "suffix": ""
                    },
                    {
                        "first": "C",
                        "middle": [],
                        "last": "Grishman",
                        "suffix": ""
                    },
                    {
                        "first": "P",
                        "middle": [],
                        "last": "Harrison",
                        "suffix": ""
                    },
                    {
                        "first": "D",
                        "middle": [],
                        "last": "Hindle",
                        "suffix": ""
                    },
                    {
                        "first": "R",
                        "middle": [],
                        "last": "Ingria",
                        "suffix": ""
                    },
                    {
                        "first": "F",
                        "middle": [],
                        "last": "Jelinek",
                        "suffix": ""
                    },
                    {
                        "first": "J",
                        "middle": [],
                        "last": "Klavans",
                        "suffix": ""
                    },
                    {
                        "first": "M",
                        "middle": [],
                        "last": "Liberman",
                        "suffix": ""
                    },
                    {
                        "first": "M",
                        "middle": [],
                        "last": "Marcus",
                        "suffix": ""
                    },
                    {
                        "first": "S",
                        "middle": [],
                        "last": "Roukos",
                        "suffix": ""
                    },
                    {
                        "first": "B",
                        "middle": [],
                        "last": "Santorini",
                        "suffix": ""
                    },
                    {
                        "first": "T",
                        "middle": [],
                        "last": "Strzalkowski",
                        "suffix": ""
                    }
                ],
                "year": 1991,
                "venue": "Proceedings of the Workshop on Speech and Natural Language, HLT '91",
                "volume": "",
                "issue": "",
                "pages": "306--311",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "E. Black, S. Abney, S. Flickenger, C. Gdaniec, C. Gr- ishman, P. Harrison, D. Hindle, R. Ingria, F. Jelinek, J. Klavans, M. Liberman, M. Marcus, S. Roukos, B. Santorini, and T. Strzalkowski. 1991. Procedure for quantitatively comparing the syntactic coverage of english grammars. In Proceedings of the Workshop on Speech and Natural Language, HLT '91, pages 306- 311, Stroudsburg, PA, USA. Association for Compu- tational Linguistics.",
                "links": null
            },
            "BIBREF2": {
                "ref_id": "b2",
                "title": "Propbank: Semantics of new predicate types",
                "authors": [
                    {
                        "first": "Claire",
                        "middle": [],
                        "last": "Bonial",
                        "suffix": ""
                    },
                    {
                        "first": "Julia",
                        "middle": [],
                        "last": "Bonn",
                        "suffix": ""
                    },
                    {
                        "first": "Kathryn",
                        "middle": [],
                        "last": "Conger",
                        "suffix": ""
                    },
                    {
                        "first": "Jena",
                        "middle": [
                            "D"
                        ],
                        "last": "Hwang",
                        "suffix": ""
                    },
                    {
                        "first": "Martha",
                        "middle": [],
                        "last": "Palmer",
                        "suffix": ""
                    }
                ],
                "year": 2014,
                "venue": "Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC'14)",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Claire Bonial, Julia Bonn, Kathryn Conger, Jena D. Hwang, and Martha Palmer. 2014. Propbank: Se- mantics of new predicate types. In Proceedings of the Ninth International Conference on Language Re- sources and Evaluation (LREC'14), Reykjavik, Ice- land, may.",
                "links": null
            },
            "BIBREF3": {
                "ref_id": "b3",
                "title": "Introduction to the conll-2005 shared task: Semantic role labeling",
                "authors": [
                    {
                        "first": "Xavier",
                        "middle": [],
                        "last": "Carreras",
                        "suffix": ""
                    },
                    {
                        "first": "Llu\u00eds",
                        "middle": [],
                        "last": "M\u00e0rquez",
                        "suffix": ""
                    }
                ],
                "year": 2005,
                "venue": "Proceedings of the Ninth Conference on Computational Natural Language Learning, CONLL '05",
                "volume": "",
                "issue": "",
                "pages": "152--164",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Xavier Carreras and Llu\u00eds M\u00e0rquez. 2005. Introduc- tion to the conll-2005 shared task: Semantic role la- beling. In Proceedings of the Ninth Conference on Computational Natural Language Learning, CONLL '05, pages 152-164, Stroudsburg, PA, USA. Associa- tion for Computational Linguistics.",
                "links": null
            },
            "BIBREF4": {
                "ref_id": "b4",
                "title": "Semi-supervised semantic role labeling using the latent words language model",
                "authors": [
                    {
                        "first": "Koen",
                        "middle": [],
                        "last": "Deschacht",
                        "suffix": ""
                    },
                    {
                        "first": "Marie-Francine",
                        "middle": [],
                        "last": "Moens",
                        "suffix": ""
                    }
                ],
                "year": 2009,
                "venue": "Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing",
                "volume": "1",
                "issue": "",
                "pages": "21--29",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Koen Deschacht and Marie-Francine Moens. 2009. Semi-supervised semantic role labeling using the la- tent words language model. In Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing: Volume 1 -Volume 1, EMNLP '09, pages 21-29, Stroudsburg, PA, USA. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF5": {
                "ref_id": "b5",
                "title": "Hownet and its computation of meaning",
                "authors": [
                    {
                        "first": "Zhendong",
                        "middle": [],
                        "last": "Dong",
                        "suffix": ""
                    },
                    {
                        "first": "Qiang",
                        "middle": [],
                        "last": "Dong",
                        "suffix": ""
                    },
                    {
                        "first": "Changling",
                        "middle": [],
                        "last": "Hao",
                        "suffix": ""
                    }
                ],
                "year": 2010,
                "venue": "Proceedings of the 23rd International Conference on Computational Linguistics: Demonstrations, COLING '10",
                "volume": "",
                "issue": "",
                "pages": "53--56",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Zhendong Dong, Qiang Dong, and Changling Hao. 2010. Hownet and its computation of meaning. In Proceed- ings of the 23rd International Conference on Compu- tational Linguistics: Demonstrations, COLING '10, pages 53-56, Stroudsburg, PA, USA. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF6": {
                "ref_id": "b6",
                "title": "WordNet: An Electronic Lexical Database",
                "authors": [
                    {
                        "first": "Christiane",
                        "middle": [],
                        "last": "Fellbaum",
                        "suffix": ""
                    }
                ],
                "year": 1998,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Christiane Fellbaum. 1998. WordNet: An Electronic Lexical Database. MIT Press.",
                "links": null
            },
            "BIBREF7": {
                "ref_id": "b7",
                "title": "Automatic labeling of semantic roles",
                "authors": [
                    {
                        "first": "Daniel",
                        "middle": [],
                        "last": "Gildea",
                        "suffix": ""
                    },
                    {
                        "first": "Daniel",
                        "middle": [],
                        "last": "Jurafsky",
                        "suffix": ""
                    }
                ],
                "year": 2002,
                "venue": "Computational Linguistics",
                "volume": "28",
                "issue": "3",
                "pages": "245--288",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Daniel Gildea and Daniel Jurafsky. 2002. Automatic la- beling of semantic roles. Computational Linguistics, 28(3):245-288.",
                "links": null
            },
            "BIBREF8": {
                "ref_id": "b8",
                "title": "Inducing chinese selectional preference based on hownet",
                "authors": [
                    {
                        "first": "Yuxiang",
                        "middle": [],
                        "last": "Jia",
                        "suffix": ""
                    },
                    {
                        "first": "Hongying",
                        "middle": [],
                        "last": "Zan",
                        "suffix": ""
                    },
                    {
                        "first": "Ming",
                        "middle": [],
                        "last": "Fan",
                        "suffix": ""
                    }
                ],
                "year": 2011,
                "venue": "Proceedings of the Seventh International Conference on Computational Intelligence and Security, CIS2011",
                "volume": "",
                "issue": "",
                "pages": "1146--1149",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Yuxiang Jia, Hongying Zan, and Ming Fan. 2011. Induc- ing chinese selectional preference based on hownet. In Proceedings of the Seventh International Conference on Computational Intelligence and Security, CIS2011, pages 1146-1149.",
                "links": null
            },
            "BIBREF9": {
                "ref_id": "b9",
                "title": "Computational models for chinese selectional preferences induction",
                "authors": [
                    {
                        "first": "Yuxiang",
                        "middle": [],
                        "last": "Jia",
                        "suffix": ""
                    },
                    {
                        "first": "Hongying",
                        "middle": [],
                        "last": "Zan",
                        "suffix": ""
                    },
                    {
                        "first": "Ming",
                        "middle": [],
                        "last": "Fan",
                        "suffix": ""
                    },
                    {
                        "first": "Shiwen",
                        "middle": [],
                        "last": "Yu",
                        "suffix": ""
                    },
                    {
                        "first": "Zhimin",
                        "middle": [],
                        "last": "Wang",
                        "suffix": ""
                    }
                ],
                "year": 2013,
                "venue": "International Journal of Advanced Intelligence",
                "volume": "5",
                "issue": "1",
                "pages": "110--119",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Yuxiang Jia, Hongying Zan, Ming Fan, , Shiwen Yu, and Zhimin Wang. 2013. Computational models for chi- nese selectional preferences induction. International Journal of Advanced Intelligence, 5(1):110-119, July.",
                "links": null
            },
            "BIBREF10": {
                "ref_id": "b10",
                "title": "Unsupervised semantic role induction via split-merge clustering",
                "authors": [
                    {
                        "first": "Joel",
                        "middle": [],
                        "last": "Lang",
                        "suffix": ""
                    },
                    {
                        "first": "Mirella",
                        "middle": [],
                        "last": "Lapata",
                        "suffix": ""
                    }
                ],
                "year": 2011,
                "venue": "Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Joel Lang and Mirella Lapata. 2011. Unsupervised se- mantic role induction via split-merge clustering. In Proceedings of the 49th Annual Meeting of the Asso- ciation for Computational Linguistics: Human Lan- guage Technologies, Portland, Oregon, USA, June. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF11": {
                "ref_id": "b11",
                "title": "Automatic retrieval and clustering of similar words",
                "authors": [
                    {
                        "first": "Dekang",
                        "middle": [],
                        "last": "Lin",
                        "suffix": ""
                    }
                ],
                "year": 1998,
                "venue": "Association for Computational Linguistics",
                "volume": "",
                "issue": "",
                "pages": "768--774",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Dekang Lin. 1998. Automatic retrieval and clustering of similar words. In Proceedings of ACL 1998, ACL '98, pages 768-774, Stroudsburg, PA, USA. Associa- tion for Computational Linguistics.",
                "links": null
            },
            "BIBREF12": {
                "ref_id": "b12",
                "title": "Plda+: Parallel latent dirichlet allocation with data placement and pipeline processing",
                "authors": [
                    {
                        "first": "Zhiyuan",
                        "middle": [],
                        "last": "Liu",
                        "suffix": ""
                    },
                    {
                        "first": "Yuzhou",
                        "middle": [],
                        "last": "Zhang",
                        "suffix": ""
                    },
                    {
                        "first": "Edward",
                        "middle": [
                            "Y"
                        ],
                        "last": "Chang",
                        "suffix": ""
                    },
                    {
                        "first": "Maosong",
                        "middle": [],
                        "last": "Sun",
                        "suffix": ""
                    }
                ],
                "year": 2011,
                "venue": "ACM Trans. Intell. Syst. Technol",
                "volume": "2",
                "issue": "3",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Zhiyuan Liu, Yuzhou Zhang, Edward Y. Chang, and Maosong Sun. 2011. Plda+: Parallel latent dirich- let allocation with data placement and pipeline pro- cessing. ACM Trans. Intell. Syst. Technol., 2(3):26:1- 26:18, May.",
                "links": null
            },
            "BIBREF13": {
                "ref_id": "b13",
                "title": "Selectional preference and sense disambiguation",
                "authors": [
                    {
                        "first": "Philip",
                        "middle": [],
                        "last": "Resnik",
                        "suffix": ""
                    }
                ],
                "year": 1997,
                "venue": "Proceedings of ACL SIGLEX Workshop on Tagging Text with Lexical Semantics",
                "volume": "",
                "issue": "",
                "pages": "52--57",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Philip Resnik. 1997. Selectional preference and sense disambiguation. In Proceedings of ACL SIGLEX Workshop on Tagging Text with Lexical Semantics, pages 52-57, Washington, D.C. ACL.",
                "links": null
            },
            "BIBREF14": {
                "ref_id": "b14",
                "title": "A latent dirichlet allocation method for selectional preferences",
                "authors": [
                    {
                        "first": "Alan",
                        "middle": [],
                        "last": "Ritter",
                        "suffix": ""
                    },
                    {
                        "first": "Oren",
                        "middle": [],
                        "last": "Etzioni",
                        "suffix": ""
                    }
                ],
                "year": 2010,
                "venue": "Proceedings of the 48th Annual Meeting of the Association for Computational Linguistics (ACL)",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Alan Ritter and Oren Etzioni. 2010. A latent dirichlet allocation method for selectional preferences. In In Proceedings of the 48th Annual Meeting of the Associ- ation for Computational Linguistics (ACL).",
                "links": null
            },
            "BIBREF15": {
                "ref_id": "b15",
                "title": "Probabilistic distributional semantics with latent variable models",
                "authors": [
                    {
                        "first": "Diarmuid",
                        "middle": [],
                        "last": "S\u00e9aghdha",
                        "suffix": ""
                    },
                    {
                        "first": "Anna",
                        "middle": [],
                        "last": "Korhonen",
                        "suffix": ""
                    }
                ],
                "year": 2014,
                "venue": "Computational Linguistics",
                "volume": "40",
                "issue": "3",
                "pages": "587--631",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Diarmuid S\u00e9aghdha and Anna Korhonen. 2014. Prob- abilistic distributional semantics with latent variable models. Computational Linguistics, 40(3):587-631, September.",
                "links": null
            },
            "BIBREF16": {
                "ref_id": "b16",
                "title": "Semantics-driven shallow parsing for chinese semantic role labeling",
                "authors": [
                    {
                        "first": "Weiwei",
                        "middle": [],
                        "last": "Sun",
                        "suffix": ""
                    }
                ],
                "year": 2010,
                "venue": "Proceedings of the ACL 2010 Conference Short Papers, ACLShort '10",
                "volume": "",
                "issue": "",
                "pages": "103--108",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Weiwei Sun. 2010. Semantics-driven shallow parsing for chinese semantic role labeling. In Proceedings of the ACL 2010 Conference Short Papers, ACLShort '10, pages 103-108, Stroudsburg, PA, USA. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF17": {
                "ref_id": "b17",
                "title": "A Bayesian approach to unsupervised semantic role induction",
                "authors": [
                    {
                        "first": "Ivan",
                        "middle": [],
                        "last": "Titov",
                        "suffix": ""
                    },
                    {
                        "first": "Alexandre",
                        "middle": [],
                        "last": "Klementiev",
                        "suffix": ""
                    }
                ],
                "year": 2012,
                "venue": "Proceedings of the Conference of the European Chapter of the Association for Computational Linguistics",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Ivan Titov and Alexandre Klementiev. 2012. A Bayesian approach to unsupervised semantic role induction. In Proceedings of the Conference of the European Chap- ter of the Association for Computational Linguistics, Avignon, France, April.",
                "links": null
            },
            "BIBREF18": {
                "ref_id": "b18",
                "title": "Labeling chinese predicates with semantic roles",
                "authors": [
                    {
                        "first": "Nianwen",
                        "middle": [],
                        "last": "Xue",
                        "suffix": ""
                    }
                ],
                "year": 2008,
                "venue": "Computational Linguistics",
                "volume": "34",
                "issue": "2",
                "pages": "225--255",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Nianwen Xue. 2008. Labeling chinese predicates with semantic roles. Computational Linguistics, 34(2):225-255.",
                "links": null
            },
            "BIBREF19": {
                "ref_id": "b19",
                "title": "More accurate tests for the statistical significance of result differences",
                "authors": [
                    {
                        "first": "Alexander",
                        "middle": [],
                        "last": "Yeh",
                        "suffix": ""
                    }
                ],
                "year": 2000,
                "venue": "Proceedings of the 18th Conference on Computational Linguistics",
                "volume": "2",
                "issue": "",
                "pages": "947--953",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Alexander Yeh. 2000. More accurate tests for the statis- tical significance of result differences. In Proceedings of the 18th Conference on Computational Linguistics - Volume 2, COLING '00, pages 947-953, Stroudsburg, PA, USA. Association for Computational Linguistics.",
                "links": null
            },
            "BIBREF20": {
                "ref_id": "b20",
                "title": "Selectional preferences for semantic role classification",
                "authors": [
                    {
                        "first": "Eneko",
                        "middle": [],
                        "last": "Be\u00f1at Zapirain",
                        "suffix": ""
                    },
                    {
                        "first": "Llu\u00eds",
                        "middle": [],
                        "last": "Agirre",
                        "suffix": ""
                    },
                    {
                        "first": "Mihai",
                        "middle": [],
                        "last": "M\u00e0rquez",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Surdeanu",
                        "suffix": ""
                    }
                ],
                "year": 2013,
                "venue": "Computational Linguistics",
                "volume": "",
                "issue": "",
                "pages": "631--663",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Be\u00f1at Zapirain, Eneko Agirre, Llu\u00eds M\u00e0rquez, and Mihai Surdeanu. 2013. Selectional preferences for seman- tic role classification. In Computational Linguistics, pages 631-663.",
                "links": null
            },
            "BIBREF21": {
                "ref_id": "b21",
                "title": "Joint inference for bilingual semantic role labeling",
                "authors": [
                    {
                        "first": "Tao",
                        "middle": [],
                        "last": "Zhuang",
                        "suffix": ""
                    },
                    {
                        "first": "Chengqing",
                        "middle": [],
                        "last": "Zong",
                        "suffix": ""
                    }
                ],
                "year": 2010,
                "venue": "Proceedings of EMNLP 2010",
                "volume": "",
                "issue": "",
                "pages": "304--314",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Tao Zhuang and Chengqing Zong. 2010. Joint inference for bilingual semantic role labeling. In Proceedings of EMNLP 2010, pages 304-314, Cambridge, MA, Oc- tober. Association for Computational Linguistics.",
                "links": null
            }
        },
        "ref_entries": {
            "FIGREF0": {
                "uris": null,
                "text": "Chinese nominal predicate translated to English verb predicate",
                "type_str": "figure",
                "num": null
            },
            "TABREF0": {
                "html": null,
                "text": "[\u9999\u6e2f \u957f\u5b98 \u8463\u5efa\u534e] [\u4eca\u5929] [\u5bf9 \u7f8e\u56fd \u57fa\u91d1\u4f1a \u53d1\u8868 \u7684 \u7ecf\u6d4e \u62a5\u544a] [\u8868\u793a] \u6b22 \u6b22 \u6b22\u8fce \u8fce \u8fce Hong Kong official Dong Jianhua today toward US foundation post economic report express welcome [AM-tmp Today], [A0 Hong Kong official Dong Jianhua] [V welcomed] [A1 the economic report released by the US foundation].",
                "content": "<table><tr><td>A0</td><td>AM-tmp</td><td>A1</td><td>Sup</td><td>V</td></tr></table>",
                "type_str": "table",
                "num": null
            },
            "TABREF1": {
                "html": null,
                "text": "",
                "content": "<table><tr><td/><td/><td/><td/><td/><td>: Topics in Chinese Gigaword</td></tr><tr><td>system</td><td>p</td><td>nominal r</td><td>f1</td><td>verb f1</td><td>all f1</td></tr><tr><td colspan=\"6\">baseline 64.71 48.20 55.25 75.53 72.08</td></tr><tr><td colspan=\"6\">SP LDA 65.70 51.27 57.59 75.93 72.74</td></tr></table>",
                "type_str": "table",
                "num": null
            },
            "TABREF2": {
                "html": null,
                "text": "SP LDA 62.70 51.03 56.27",
                "content": "<table><tr><td colspan=\"4\">: Chinese PropBank 1.0 results</td><td/></tr><tr><td>sections</td><td>system</td><td>p</td><td>r</td><td>f1</td></tr><tr><td colspan=\"5\">Sinorama baseline 37.58 25.10 30.10</td></tr><tr><td>nominal</td><td colspan=\"4\">SP LDA 39.72 27.36 32.40</td></tr><tr><td>verb</td><td colspan=\"4\">baseline 67.13 50.37 57.55 SP LDA 67.56 50.59 57.86</td></tr><tr><td>4051-</td><td colspan=\"4\">baseline 62.01 50.74 55.81</td></tr><tr><td>4411 (verb)</td><td/><td/><td/><td/></tr></table>",
                "type_str": "table",
                "num": null
            },
            "TABREF3": {
                "html": null,
                "text": "Chinese PropBank 3.0 out-of-genre results",
                "content": "<table/>",
                "type_str": "table",
                "num": null
            },
            "TABREF5": {
                "html": null,
                "text": "",
                "content": "<table><tr><td/><td colspan=\"4\">: Chinese SRL comparison 6</td></tr><tr><td>system</td><td>p</td><td>r</td><td>f1</td><td>error \u2206</td></tr><tr><td>SwiRL</td><td colspan=\"3\">79.7 70.9 75.0</td><td/></tr><tr><td colspan=\"5\">Zapirain 2013 80.0 71.3 75.4 \u22121.60%</td></tr><tr><td>baseline</td><td colspan=\"3\">82.59 77.27 79.84</td><td/></tr><tr><td>SP LDA</td><td colspan=\"4\">82.96 77.52 80.15 \u22121.54%</td></tr></table>",
                "type_str": "table",
                "num": null
            },
            "TABREF6": {
                "html": null,
                "text": "",
                "content": "<table><tr><td>: English SRL comparison (CoNLL-2005 WSJ)</td></tr><tr><td>CoNLL-2005 development set. Compared to Zapi-</td></tr><tr><td>rain et al. (2013) (table</td></tr></table>",
                "type_str": "table",
                "num": null
            }
        }
    }
}