Báo cáo khoa học: "Attacking Parsing Bottlenecks with Unlabeled Data and Relevant Factorizations" pdf

... Association for Computational Linguistics Attacking Parsing Bottlenecks with Unlabeled Data and Relevant Factorizations Emily Pitler Computer and Information Science University of Pennsylvania Philadelphia, ... 84.3 dpo3 (Grand+Sib) 93.21 44.8 89.6 86.9 dpo3 +Unlabeled (Edges) 93.12 43.6 85.3 87.0 dpo3 +Unlabeled (Sib) 93.15 43.7 85.5 86.8 dpo3 +Unlabeled (Grand) 93.55 46.1...

Ngày tải lên: 23/03/2014, 14:20

... 6.8% serve supply with food 42.6% (verb) hold an office 33.6% function as something 16% provide a service 7.8% 3 Experiments and Evaluation 3.1 Test data We constructed four datasets from hand-tagged ... (Pantel and Lin, 2002; Sch¨utze, 1998), there are other related efforts on word sense discrimination (Dorow and Widdows, 2003; Fukumoto and Suzuki, 1999; Pedersen and Bruce, 1997...

Ngày tải lên: 20/02/2014, 16:20

... Dependency Treebank) and English (CoNLL-2009) datasets. The scores in brackets are achieved with gold-standard POS tagging. since it is overt in all the other forms, tenses and moods of the verb. ... working with gold-standard POS tags, which suggests that do- main difficulties harm POS tagging and parsing as well. Regarding the two last subcorpora, the com- positions consist of ve...

Ngày tải lên: 17/03/2014, 22:20

... to spoken dialogue systems that includes rule-based and trainable dialogue managers, spoken language understanding and generation modules, and a compre- hensive dialogue system architecture. ... simula- tions. To optimize Q and populate the policy with ex- pected values, the learner needs to explore un- tried actions (system moves) to gain more expe- riences, and combine this wit...

Ngày tải lên: 23/03/2014, 17:20

... Textual Parallelisms with edit distance and similarity degree Marie Gu ´ egan and Nicolas Hernandez LIMSI-CNRS Universit´e de Paris-Sud, France guegan@aist.enst.fr | hernandez@limsi.fr Abstract Detection ... a string editing distance (Wagner and Fischer, 1974) and a tree editing distance 1 (Zhang and Shasha, 1989). Section 4 discusses and evaluates these methods and their rele...

Ngày tải lên: 24/03/2014, 03:20

... ASR and rule-based MT coupling: a) First-best b) N-best list c) N-best word graph. While integrating the SR system with the rule- based MT system, this study uses word graphs and chart parsing with ... from standard theory and algorithms on FSMs. In the converted FSM, non-determinism is removed and it is minimized by eliminating redun- dant nodes and arcs. Next, the chart is i...

Ngày tải lên: 31/03/2014, 20:20

... only one parse tree with highest possibility. During the parsing process, the parser keeps the unexpanded active nodes in a heap, and always expands the active node with the best probability. ... co-occurence of words a and b with the indepen- dent probabilities of occurrence of a and b (Church and Hanks, 1990). P(a, b) I(a, b) = log P(a)P(b) where the probabilities...

Ngày tải lên: 31/03/2014, 21:20

... k 2 sentiment words and topic words with highest scores as candidates. 5: Construct a bipartite graph between sentiment and topic words on D l T and the k 2 sentiment- and topic- word can- didates, and calculate ... scores  S 1 and  S 3 of the k 2 sentiment and topic word candidates using Eqs. (5) and (6) iteratively. 7: Select k 1 new sentiment words and k 1 new topic w...

Ngày tải lên: 19/02/2014, 19:20

... be directly compared with constituents proposed within the HPSG parse. Consider the sentence "Robin promised to come at noon", with the following context: Sandy: "I guess we ... Research Funded by The Ohio State Center for Cognitive Science and The Ohio State Departments of Computer and Information Science and Linguistics grammar (using compiled knowledge) wh...

Ngày tải lên: 20/02/2014, 21:20

... ambiguous word in accordance with frequency information, with- out considering the extent to which the features co- occur with one another. Gale, Church and Yarowsky ([10]) and Yarowsky ([29]) formally ... efficiency, and providing an understanding of the data. Further, different types of variables, such as class-based and collocation-specific ones, can be used in combinat...

Ngày tải lên: 23/03/2014, 20:21