Efficient and Interpretable Grammatical Error Correction with Mixture of Experts
This work addresses efficiency and interpretability issues in GEC for NLP practitioners, representing an incremental improvement over existing combination methods.
The paper tackles the computational inefficiency of combining multiple grammatical error correction (GEC) models by proposing MoECE, a mixture-of-experts model that achieves the performance of T5-XL with three times fewer effective parameters and provides interpretable error type identification.
Error type information has been widely used to improve the performance of grammatical error correction (GEC) models, whether for generating corrections, re-ranking them, or combining GEC models. Combining GEC models that have complementary strengths in correcting different error types is very effective in producing better corrections. However, system combination incurs a high computational cost due to the need to run inference on the base systems before running the combination method itself. Therefore, it would be more efficient to have a single model with multiple sub-networks that specialize in correcting different error types. In this paper, we propose a mixture-of-experts model, MoECE, for grammatical error correction. Our model successfully achieves the performance of T5-XL with three times fewer effective parameters. Additionally, our model produces interpretable corrections by also identifying the error type during inference.