2 datasets found

Propose a system for grammatical error correction using a masked language model. Introduce a framework consisting of two tasks: tagging, which chooses and arbitrarily reorders a...

The WMT’14 English-German and English-French benchmarks, CNN/DailyMail dataset, and CONLL dataset were used for machine translation, abstractive summarization, and grammar error...