Paper: Constrained Grammatical Error Correction using Statistical Machine Translation

ACL ID W13-3607
Title Constrained Grammatical Error Correction using Statistical Machine Translation
Venue International Conference on Computational Natural Language Learning
Session shared task
Year 2013
Authors

This paper describes our use of phrase- based statistical machine translation (PB- SMT) for the automatic correction of er- rors in learner text in our submission to the CoNLL 2013 Shared Task on Gram- matical Error Correction. Since the lim- ited training data provided for the task was insufficient for training an effective SMT system, we also explored alternative ways of generating pairs of incorrect and correct sentences automatically from other existing learner corpora. Our approach does not yield particularly high perfor- mance but reveals many problems that re- quire careful attention when building SMT systems for error correction.