Intricacies of Collins’ Parsing Model

By D. Bikel. International Conference on Computational Logic, 2004

307 citations recorded by the source corpus.

This article documents a large set of heretofore unpublished details Collins used in his parser, such that, along with Collins' (1999) thesis, this article contains all information necessary to duplicate Collins' benchmark results. Indeed, these as-yet-unpublished details account for an 11 relative increase in error from an implementation including all details to a clean-room implementation of Collins' model. We also show a cleaner and equally well-performing method for the handling of punctuation and conjunction and reveal certain other probabilistic oddities about Collins' parser. We not only analyze the effect of the unpublished details, but also reanalyze the effect of certain well-known details, revealing that bilexical dependencies are barely used by the model and that head choice is not nearly as important to overall parsing performance as once thought. Finally, we perform experiments that show that the true discriminative power of lexicalization appears to lie in the fact that unlexicalized syntactic structures are generated conditioning on the headword and its part of speech.

Subfields: Artificial Intelligence, Language and Linguistics

View Paper: https://www.semanticscholar.org/paper/0606291dae96446e812ea8f09d9fbdc6acc3ec37

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Language and Linguistics

Citations in the corpus

Developing a Robust Part-of-Speech Tagger for Biomedical Text (2005, background): “The precision of POS tagging not only directly affects the performance of pattern-based approaches but also influences the accuracy of parsing which in general uses the POS tags on the words as part of the input [1,2].”
The Necessity of Syntactic Parsing for Semantic Role Labeling (2005, methodology): “We use the Collins’ parser implemented by Bikel (2004).”
Constituent Parsing with Incremental Sigmoid Belief Networks (2007, methodology): “Table 1 lists the results of the NN approximation and the MF approximation, along with results of dif-2 ferent generative and discriminative parsing meth-ods (Bikel, 2004; Taskar et al., 2004; Turian and Melamed, 2006; Charniak, 2000) evaluated in the same experimental setup.”
Incremental Sigmoid Belief Networks for Grammar Learning (2010, methodology): “…and the IMF approximation,7 along with results of different generative and discriminative parsing methods evaluated in the same experimental setup (Bikel, 2004; Taskar et al., 2004; Turian et al., 2006; Charniak, 2000; Finkel et al., 2008).8 The IMF model improves over the baseline NN…”
The Importance of Syntactic Parsing and Inference in Semantic Role Labeling (2008, methodology): “Formally speaking, the argument classifiers attempt to assign 3 We use the Collins parser implemented by Bikel (2004). labels to a set of arguments, S1:M, indexed from 1 toM.”
Joint Parsing and Semantic Role Labeling (2005, methodology): “The k -best list of parses is generated using Dan Bikel’s (2004) implementation of Michael Collins’ parsing model.”
Better k-best Parsing (2005, methodology): “For probabilistic parsing, we implemented Algorithms 0, 1, and 3 on top of a widely-used parser (Bikel, 2004) and conducted experiments on parsing efficiency and the quality of the k-best-lists.”
Online Large-Margin Training of Dependency Parsers (2005, methodology): “We compared our system to the Bikel re-implementation of the Collins parser (Bikel, 2004; Collins, 1999) trained with the same head rules of our system.”