Rule-based Coreference Resolution System
A Rule-based Coreference Resolution System is a Coreference Resolution System that is based on set of Coreference Rules.
- AKA: Heuristics-based Coreference Resolution System.
- Context:
- It can solve a Rule-based Coreference Resolution Task by implementing a Rule-based Coreference Resolution Algorithm.
- Example(s):
- COCKTAIL (Harabagiu et al., 2001),
- CogNIAC (Baldwin, 1997),
- Deterministic Rule-based High-Precision Entity-Centric Coreference Resolution System (Lee et al., 2013),
- An Easy-First Coreference Resolution System (Stoyanov & Eisner, 2012),
- A Multi-Pass Sieve Coreference Resolution System (Raghunathan et al., 2010).
- …
- Counter-Example(s):
- An AllSingleton Coreference Resolution System,
- An OneCluster Coreference Resolution System,
- A Better Baseline Coreference Resolution System,
- A Centroid-based Joint Canonicalization-Coreference Resolution System,
- A Ranking-based Coreference Resolution System,
- A Stacking-based Coreference Resolution System,
- A Graph-based Coreference Resolution System,
- A Classification-based Coreference Resolution System such as:
- See: Entity Mention Normalization System, Natural Language Processing System, Information Extraction System.
References
2015
- (Sawhney & Wang, 2015) ⇒ Kartik Sawhney, and Rebecca Wang. (2015). “Coreference Resolution.”
- QUOTE: In this section, we discuss our hand-written rules used in the rule-based coreference system and their observed impact. Our implementation is heavily inspired by "A Multi-Pass Sieve for Coreference Resolution", Raghunathan et al. In particular, we adopt a similar multi-sieve approach, transitioning from high precision to high recall. The training phase collects statistics on head word matches to be able to implement the simple heard word matching algorithm described in the “better baseline” model above. The following is a discussion of the different passes.
Exact match (pass1)...
Head matching based on training data (pass2)...
Strict head matching (pass3) ...
Variants of simple head matching (passes4 and 5) ...
Relaxed head matching (pass6)...
Nominal coreference using gender, number, speaker and lemmas (pass7)...
Hobbs' algorithm (pass8) ...
Pronominal coreference using gender, number and speaker (pass9)...
- QUOTE: In this section, we discuss our hand-written rules used in the rule-based coreference system and their observed impact. Our implementation is heavily inspired by "A Multi-Pass Sieve for Coreference Resolution", Raghunathan et al. In particular, we adopt a similar multi-sieve approach, transitioning from high precision to high recall. The training phase collects statistics on head word matches to be able to implement the simple heard word matching algorithm described in the “better baseline” model above. The following is a discussion of the different passes.
2013
- (Lee et al., 2013) ⇒ Heeyoung Lee, Angel Chang, Yves Peirsman, Nathanael Chambers, Mihai Surdeanu, and Dan Jurafsky. (2013). “Deterministic Coreference Resolution based on Entity-centric, Precision-ranked Rules.” In: Computational Linguistics Journal, 39(4). doi:10.1162/COLI_a_00152
- QUOTE: Rule-based models like Lappin and Leass (1994) were a popular early solution to the subtask of pronominal anaphora resolution. Rules are easy to create and maintain and error analysis is more transparent. But early rule-based systems relied on hand-tuned weights and were not capable of global inference, two factors that led to poor performance and replacement by machine learning.
We propose a new approach that brings together the insights of these modern supervised and unsupervised models with the advantages of deterministic, rule-based systems. We introduce a model that performs entity-centric coreference, where all mentions that point to the same real-world entity are jointly modeled, in a rich feature space using solely simple, deterministic rules. Our work is inspired both by the seminal early work of Baldwin (1997), who first proposed that a series of high-precision rules could be used to build a high-precision, low-recall system for anaphora resolution, and by more recent work that has suggested that deterministic rules can outperform machine learning models for coreference (Zhou and Su 2004; Haghighi and Klein 2009) and for named entity recognition (Chiticariu et al. 2010).
- QUOTE: Rule-based models like Lappin and Leass (1994) were a popular early solution to the subtask of pronominal anaphora resolution. Rules are easy to create and maintain and error analysis is more transparent. But early rule-based systems relied on hand-tuned weights and were not capable of global inference, two factors that led to poor performance and replacement by machine learning.
2012
- (Stoyanov & Eisner, 2012) ⇒ Veselin Stoyanov, and Jason Eisner. (2012). “Easy-first Coreference Resolution.” In: Proceedings of COLING 2012.
- QUOTE: We propose a coreference resolution approach that like Raghunathan et al. (2010) aims to consider global consistency while performing fast and deterministic greedy search. Similar to Raghunathan et al. (2010), our algorithm operates by making the easy (most confident) decisions first. It builds up coreference clusters as it goes and uses the information from these clusters in the form of features to make later decisions. However, while Raghunathan et al. (2010) use hand-written rules for their system, we learn feature weights from training data.
2010
- (Raghunathan et al., 2010) ⇒ Karthik Raghunathan, Heeyoung Lee, Sudarshan Rangarajan, Nathanael Chambers, Mihai Surdeanu, Dan Jurafsky, and Christopher Manning. (2010). “A Multi-pass Sieve for Coreference Resolution.” In: Proceedings of the 2010 Conference on Empirical Methods in Natural Language Processing.
- QUOTE: We propose an unsupervised sieve-like approach to coreference resolution that addresses these issues. The approach applies tiers of coreference models one at a time from highest to lowest precision. Each tier builds on the entity clusters constructed by previous models in the sieve, guaranteeing that stronger features are given precedence over weaker ones. Furthermore, each model’s decisions are richly informed by sharing attributes across the mentions clustered in earlier tiers. This ensures that each decision uses all of the information available at the time. We implemented all components in our approach using only deterministic models. All our components are unsupervised, in the sense that they do not require training on gold coreference links.
2001
- (Harabagiu et al., 2001) ⇒ Sanda M. Harabagiu, Rǎzvan C. Bunescu, and Steven J. Maiorano. (2001). “Text and Knowledge Mining for Coreference Resolution.” In: Proceedings of the second meeting of the North American Chapter of the Association for Computational Linguistics on Language technologies. doi:10.3115/1073336.1073344
- QUOTE: To analyze coreference data we use a corpus of annotated texts. To devise minimalist coreference resolution rules we consider (1) strong indicators of cohesion, such as repetitions, name aliases or appositions; and (2) gender, number and class agreements. WordNet (Miller 1995), the vast semantic knowledge base, provides suplementary knowledge in the form of semantic consistency between coreferring nouns. Additional semantic consistency knowledge is generated by bootstrapping when our coreference resolution system, COCKTAIL [1], processes new texts.
1997
- (Baldwin, 1997) ⇒ Breck Baldwin. (1997). “CogNIAC: High Precision Coreference with Limited Knowledge and Linguistic Resources.” In: Proceedings of a Workshop on Operational Factors in Practical, Robust Anaphora Resolution for Unrestricted Texts.
- QUOTE: CogNIAC is a pronoun resolution engine designed around the assumption that there is a sub-class of anaphora that does not require general purpose reasoning. The kinds of information CogNIAC does require includes: sentence detection, part-of-speech tagging, simple noun phrase recognition, basic semantic category information like, gender, number, and in one configuration, partial parse trees.
(...) The method of resolving pronouns within CogNIAC works as follows: Pronouns are resolved left-to-right in the text. For each pronoun, the rules are applied in the presented order. For a given rule, if an antecedent is found, then the appropriate annotations are made to the text and no more rules are tried for that pronoun, otherwise the next rule is tried. If no rules resolve the pronoun, then it is left unresolved. These rules are individually are high precision rules, and collectively they add up to reasonable recall. The precision is 97% (121/125) and the recall is 60% (121/201) for 198 pronouns of training data.
- QUOTE: CogNIAC is a pronoun resolution engine designed around the assumption that there is a sub-class of anaphora that does not require general purpose reasoning. The kinds of information CogNIAC does require includes: sentence detection, part-of-speech tagging, simple noun phrase recognition, basic semantic category information like, gender, number, and in one configuration, partial parse trees.
- ↑ COCKTAIL is a pun on COGNIAC, because COCKTAIL uses multiple coreference resolution rules corresponding to different forms of coreference, blended together in a single system