Please join the Responsible AI Systems & Experiences (RAISE) group at the University of Washington for the talk “Towards Trustworthy Evaluation Frameworks in Natural Language Generation” by Mona Diab. 

Mona Diab is a Research Scientist with Facebook AI Research and a full Professor of CS at the George Washington University. Please see the attached flyer for more information. 

Abstract: In this talk, I propose a framework for trustworthy evaluations for NLP in general with an application to natural language generation in particular. I posit that trustworthiness emerges from *both* trusting the process by which we build technologies as well as trusting the end product. I delve into the dimensions of such frameworks and the impact of devising practical metrics that correlate well with users’ perception of performance. I will speak about two evaluation metrics that optimize for faithfulness in abstractive summarization and for human evaluation for machine translation. 

Zoom link: https://washington.zoom.us/j/94636255672 

More info: https://www.raise.uw.edu/