Fetching the paper…

Toward Human-Like Evaluation for Natural Language Generation with Error Analysis · Around