Maarten Sap

Maarten Sap

I am a Postdoc/Young Investigator at the Allen Institute for AI (AI2), working on project Mosaic, and will be starting as an assistant professor at CMU's LTI department. My research focuses on endowing NLP systems with social intelligence and social commonsense, and understanding social inequality and bias in language.

I received my PhD from the University of Washington where I was advised by Noah Smith and Yejin Choi, and have interned at AI2 working on social commonsense reasoning, and at Microsoft Research working on deep learning models for understanding human cognition.
[bio for talks]

July 2021 update๐Ÿ‘จ๐Ÿผโ€๐ŸŽ“: I successfully defended my PhD thesis titled Positive AI with Social Commonsense Models (read the thesis here, or watch the recording here). Thanks to my advisors, committee, and everyone who attended!

May 2021 update๐Ÿฅณ: I will be joining CMU's LTI department as an assistant professor๐Ÿ‘จ๐Ÿผโ€๐Ÿซin Fall 2022. If you wish to work with me, see the "contact" page. Before starting there, I will be a postdoc at AI2 on project Mosaic ๐Ÿ‘จ๐Ÿผโ€๐Ÿ”ฌ starting Fall 2021.



We create an unsupervised model for controllable debiasing to rewrite and debias how characters are portrayed in sentences.

Details Watch EMNLP talk Read paper
Social Bias Frames

We introduce a new formalism to capture implications of social bias in language, and 150k structured annotations on social media posts.

Details Watch ACL talk Read paper
Social IQa

We create a large-scale benchmark for social commonsense reasoning that is challenging for modern NLP models.

Details Read paper
Racial Bias in Toxic Language Detection

We quantify racial bias in hate speech datasets and algorithms, and study how this bias arises in annotations.

Watch ACL talk Read paper

Introducing a knowledge graph for machine commonsense covering if-then inferential knowledge aroung everyday situations.

Browse Demo Read paper

We create a model and knowledge graph that enables commonsense inference on intents and reactions in relation to events.

Browse Demo Read paper
Social Commonsense in Short Stories

We create a new annotation framework of simple commonsense stories enabling reasoning about the mental states of its characters.

Browse stories Read paper
Sounding Board

Winner of the 2017 Alexa Prize to further conversational AI, our approach is user-centric and content-driven.

Learn more Read paper
Power & Agency in movies

We create connotation frames of power and agency and use them to analyze gender bias in movies.

Explore movies Read paper

External links

Twitter Profile Google Scholar Profile Semantic Scholar Profile