System Combination for Grammatical Error Correction

Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 951–962, October 25-29, 2014, Doha, Qatar. c⃝2014 Association for Computational Linguistics System Combination for Grammatical Error Correction Raymond Hendy Susanto Peter Phandi Hwee Tou Ng Department of Computer Science National University of Singapore 13 Computing Drive, Singapore 117417 {raymondhs,peter-p,nght}@comp.nus.edu.sg Abstract Different approaches to high-quality grammatical error correction have been proposed recently, many of which have their own strengths and weaknesses. Most of these approaches are based on classi- fication or statistical machine translation (SMT). In this paper, we propose to com- bine the output from a classification-based system and an SMT-based system to improve the correction quality. We adopt the system combination technique of Heafield and Lavie (2010). We achieve an F0.5 score of 39.39% on the test set of the CoNLL-2014 shared task, outperforming the best system in the shared task. 1 Introduction Grammatical error correction (GEC) refers to the task of detecting and correcting grammatical er- rors present in a text written by a second language learner. For example, a GEC system to correct English promises to benefit millions of learners around the world, since it functions as a learning aid by providing instantaneous feedback on ESL writing. Research in this area has attracted much interest recently, with four shared tasks organized in the past several years: Helping Our Own (HOO) 2011 and 2012 (Dale and Kilgarriff, 2010; Dale et al., 2012), and the CoNLL 2013 and 2014 shared tasks (Ng et al., 2013; Ng et al., 2014). Each shared task comes with an annotated corpus of learner texts and a benchmark test set, facilitating further re- search in GEC. Many approaches have been proposed to de- tect and correct grammatical errors. The most dominant approaches are based on classification (a set of classifier modules where each module ad- dresses a specific error type) and statistical ma- chine translation (SMT) (formulated as a transla- tion task from “bad” to “good” English). Other ap- proaches combine the classification and SMT ap- proaches, and often have some rule-based compo- nents. Each approach has its own strengths and weak- nesses. Since the classification approach is able to focus on each individual error type using a sep- arate classifier, it may perform better on an er- ror type where it can build a custom-made classi- fier tailored to the error type, such as subject-verb agreement errors. The drawback of the classifica- tion approach is that one classifier must be built for each error type, so a comprehensive GEC sys- tem will need to build many classifiers which com- plicates its design. Furthermore, the classification approach does not address multiple error types that may interact. The SMT approach, on the other hand, natu- rally takes care of interaction among words in a sentence as it attempts to find the best overall cor- rected sentence. It usually has a better coverage of different error types. The drawback of this ap- proach is its reliance on error-annotated learner data, which is …

Content truncated to 500 words