Paper: The AMU System in the CoNLL-2014 Shared Task: Grammatical Error Correction by Data-Intensive and Feature-Rich Statistical Machine Translation

ACL ID W14-1703
Title The AMU System in the CoNLL-2014 Shared Task: Grammatical Error Correction by Data-Intensive and Feature-Rich Statistical Machine Translation
Venue International Conference on Computational Natural Language Learning
Session shared task
Year 2014
Authors

Statistical machine translation toolkits like Moses have not been designed with gram- matical error correction in mind. In or- der to achieve competitive results in this area, it is not enough to simply add more data. Optimization procedures need to be customized, task-specific features should be introduced. Only then can the decoder take advantage of relevant data. We demonstrate the validity of the above claims by combining web-scale language models and large-scale error-corrected texts with parameter tuning according to the task metric and correction-specific fea- tures. Our system achieves a result of 35.0% F 0.5 on the blind CoNLL-2014 test set, ranking on third place. A similar sys- tem, equipped with identical models but without tuned parameters and specialized features, stagnates a...