Báo cáo khoa học: "Unsupervised Event Coreference Resolution with Rich Linguistic Features" potx

... Computational Linguistics, pages 1412–1422,Uppsala, Sweden, 11-16 July 2010.c2010 Association for Computational LinguisticsUnsupervised Event Coreference Resolution with Rich Linguistic FeaturesCosmin ... trained on.Moreover, since event coreference resolution is acomplex task that involves exploring a rich set of linguistic features, annotating a large corpus with event coreference information ... 1.2.2 Linguistic FeaturesThe events representing coreference clusters of event mentions are characterized by a large set of linguistic features. To compute an accurate event distribution for event...

Báo cáo khoa học: "Text-level Discourse Parsing with Rich Linguistic Features" pdf

... for Computational Linguistics, pages 60–68,Jeju, Republic of Korea, 8-14 July 2012.c2012 Association for Computational LinguisticsText-level Discourse Parsing with Rich Linguistic FeaturesVanessa ... discourseparser with state-of-the-art performance. We signif-icantly improve the performance of HILDA’s tree-building step (introduced in Section 5.1 below) byincorporating rich linguistic features ... we improve the over-all performance of these two classiﬁers by incorpo-rating rich linguistic features, together with appro-priate feature selection. We also explore how thesetwo classiﬁers...

Báo cáo khoa học: "Unsupervised Lexicon-Based Resolution of Unknown Words for Full Morphological Analysis" doc

... 728–736,Columbus, Ohio, USA, June 2008.c2008 Association for Computational LinguisticsUnsupervised Lexicon-Based Resolution of Unknown Words for FullMorphological AnalysisMeni Adler and Yoav ... WorkMost of the work that dealt with unknowns in the lastdecade focused on unknown tokens (OOV). A naiveapproach would assign all possible analyses for eachunknown token with uniform distribution, ... distribution, and con-tinue disambiguation on the basis of a learned model with this initial distribution. The performance of atagger with such a policy is actually poor: there aredozens of tags in the...

Báo cáo khoa học: "Optimization in Coreference Resolution Is Not Needed: A Nearly-Optimal Algorithm with Intensional Constraints" ppt

... LinguisticsOptimization in Coreference Resolution Is Not Needed:A Nearly-Optimal Algorithm with Intensional ConstraintsManfred Klenner &´Etienne AilloudComputational LinguisticsZurich ... of coreference resolution. Then we go into the de-tails of our Balas adaptation and provide empiri-cal evidence for our central claim—that optimiza-tion search can already be stopped (without ... (Telljohannet al., 2005; Naumann, 2006) coreference corpusis used. The T¨uBa is a treebank (1,100 Germannewspaper texts, 25,000 sentences) augmented with coreference annotations2. In total,...

Tài liệu Báo cáo khoa học: "Unsupervised Semantic Role Induction with Global Role Ordering" doc

... for Computational Linguistics, pages 145–149,Jeju, Republic of Korea, 8-14 July 2012.c2012 Association for Computational LinguisticsUnsupervised Semantic Role Induction with Global Role OrderingNikhil ... argument instances.We chose a uniform Dirichlet prior with concentra-tion parameter as 0.1 for all the model parametersin Algorithm 1 (set roughly, without optimization1).50 training iterations ... p dofor all voice vc ∈ {active, passive} dodraw θorderp,vc∼ Dirichlet(αorder)for all interval I dodraw θSRp,I∼ Dirichlet(αSR)for all adjacency adj ∈ {0, 1} dodraw θST OPp,I,adj∼...

Tài liệu Báo cáo khoa học: "Kernel-Based Pronoun Resolution with Structured Syntactic Knowledge" pptx

... Computational Linguistics and 44th Annual Meeting of the ACL, pages 41–48,Sydney, July 2006.c2006 Association for Computational LinguisticsKernel-Based Pronoun Resolution with Structured ... syntactic fea-ture incorporated with kernels can signiﬁcantlyimprove the resolution performance (by 5%∼8%in the success rates), and is reliably effective forthe pronoun resolution task.The remainder ... pattern with the pronominal anaphor.Table 1: Feature set for the baseline pronoun res-olution systemsalience measures have to be assigned manually.Luo and Zitouni (2005) proposed a coreference resolution...

Báo cáo khoa học: "Unsupervised Argument Identiﬁcation for Semantic Role Labeling" potx

... bear a semanticconnection with that predicate, they consequentlytend to collocate with it.We collect collocation statistics from a largecorpus, which we annotate with parse trees andPOS tags. ... constituentsthat appear to be weakly collocating with the pred-icate in question. Since a predicate greatly re-stricts the type of arguments with which it mayappear (this is often referred ... arguments with which it is likely to collocate.3.2 Clause detection stageThe main idea behind this stage is the observationthat most of the arguments of a predicate are con-tained within the...

Báo cáo khoa học: "Unsupervised Learning of Dependency Structure for Language Modeling" potx

... built with and without PW. In Table 2, the PW- prefix refers to the DLMs with PW = 0.5, and the DLMs without PW- prefix refers to DLMs with PW = 0. For both DLM_1 and DLM_2, models with the ... useful information for language modeling, the most linguistically related word in the previous context may come in various linguistic relations with the word being predicted, not limited to syntactic ... relations between headwords of each phrase in a sentence. To deal with the first obstacle mentioned above, we approximate long-distance linguistic dependency by a model that is similar to a skipping...

Báo cáo khoa học: "Unsupervised Learning of Narrative Schemas and their Participants" potx

... represent-ing events, and for each verb, the grammatical roleﬁlled by the shared protagonist.An event is a verb together with its constellationof arguments. An event slot is a tuple of an event and ... incorrectly favor (ﬂy X) because it isobserved during training with all ﬁve event slots,although not frequently with any one of them. An event slot like (charge X) is much more plausible,but is ... acrossall event slot pairs in a criminal scenario chain.This chain will be part of a larger narrativeschema, described in section 3.4.3.3 Event Slot Similarity with ArgumentsWe now formalize event...

Báo cáo khoa học: "Unsupervised Semantic Role Induction via Split-Merge Clustering" potx

... over syntactically analyzed sentenceswithout access to any data annotated with semanticroles. Argument identiﬁcation is carried out througha small set of linguistically-motivated rules, whereasrole ... identiﬁed argu-ments (auto/auto), gold parses with automatic argu-ments (gold/auto), automatic parses with gold argu-ments (auto/gold) and gold parses with gold argu-ments (gold/gold). Bold-face ... Association for Com-putational Linguistics HLT 2009: Short Papers, pages13–16, Boulder, Colorado.1126Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics, pages 1117–1126,Portland,...

Xem thêm