Matthew Lease

This paper examines how data affects the evaluation of NLP models.

New strategies aim to create respectful dialogue in online conversations.

A new method promotes fairness in identifying targeted toxic language across demographic groups.