Parser available at http://nlp.cs.berkeley.edu To conclude, ... The dynamic programming results are significantly better than the reranking results, ...
Learning and Inference for Hierarchically Split PCFGs Slav Petrov and Dan Klein The Game of Designing a Grammar Annotation refines base treebank symbols to improve ...
Learning accurate, compact, and interpretable tree annotation ... Compact Grammar. Very tedious. Allocates splits where needed. Captures many features ...
The Game of Designing a Grammar. Annotation refines base treebank symbols to improve ... [Goodman 97, Charniak&Johnson 05] Coarse grammar. NP ... VP ...
State-of-the-art accuracy for English and Chinese. 9/2/09. Qin Iris Wang (University of Alberta) ... A vector of feature weights. A vector of features. 9/2/09 ...
Cognitive Sciences ... Rooted in cognitive/psycholinguistic accounts of the ... Discourse planners to select content from data and knowledge bases and organise ...
A single symbol for all NPs. The difference between. Sbj-NP and ... We cannot use DP like the F-B algo. to. solve (1) (NP-hard) ... (1) Parsing by Approximation ...
Synchronous TAG (STAG) Multi-component TAG (MCTAG) TAG basics ... Many extensions: LTAG, STAG, ... Statistical approach. Modeling. Choose the objective function ...
Learning and Inference. for Hierarchically Split PCFGs. Slav ... a Grammar. Annotation refines base treebank symbols to improve statistical fit of the grammar ...
But: Hypotheses are recombined - many good translations don't reach the sentence end ... Discriminative training, to retrain given models: e.g. perceptron learning; ...