Kunal Handa

I am an MSc by Research student at the University of Oxford advised by Yarin Gal. Previously, I was a linguistics and computer science undergrad at Brown University.

I {am, have been} fortunate to work on learning under ambiguity with Noah Goodman, NLP in education with Dora Demszky and Diyi Yang, interpretability of embodied AI with Ellie Pavlick and Chen Sun, and the Peekbank Project with Michael C. Frank.

I am broadly interested in understanding the fundamentals of learning in humans and machines: 1) applying the principles of human learning to ML—building ML models that can learn in the real world—and 2) using these models to supplement how humans learn.

You can contact me at: kunal [underscore] handa [at] alumni [dot] brown [dot] edu

Currently,

I'm exploring how to improve LMs' preference-learning abilities with Alex Tamkin, Belinda Li, Ellie Pavlick, Noah Goodman, and Jacob Andreas.

I'm developing a novel formulation for uncertainty in LMs with Yarin Gal.

Publications

Conference Papers

“Mistakes Help Us Grow”: Facilitating and Evaluating Growth Mindset Supportive Language in Classrooms
Kunal Handa, Margarett Clapper, Jessica Boyle, Rose E Wang, Diyi Yang, David S Yeager, Dorottya Demszky. In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), 2023.

Emergence of Abstract State Representations in Embodied Sequence Modeling
Tian Yun*, Zilai Zeng*, Kunal Handa, Ashish V Thapliyal, Bo Pang, Ellie Pavlick, Chen Sun. In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), 2023.

Task Ambiguity in Humans and Language Models
Alex Tamkin*, Kunal Handa** denotes equal contribution., Avash Shrestha, Noah Goodman. In the International Conference on Learning Representations (ICLR), 2023.

Journal Articles

Peekbank: An open, large-scale repository for developmental eye-tracking data of children's word recognition
Martin Zettersten... Kunal Handa... & Michael C Frank.outside of the position of the first and the last author, authorship position was determined by sorting authors' last names in reverse alphabetical order. In Behavior Research Methods (BRM), 2022.

Other Writing

The Role of Technology in Elections: The Voyage of Voters' Data
Kunal Handa. In Conduit, the Brown University Computer Science Annual Magazine, Volume 32, 2022.

Trying to Give a Shit About "Give a Shit": A Compositional Semantics Perspective
Kunal Handa

Teaching

At Brown, I served as the Socially-Responsible Computing Teaching Assistant for CS1470: Deep Learning. I designed contentsome of which is available via the course's website on deep learning's potential societal harms and conducted exercises that examined ethical frameworks, the cyclical nature of language models' biases, and the pros and cons of regulating ML advancements.

I also tutor incarcerated and previously incarcerated individuals in HiSET test preparation and essay writing. Intermittently, I teach group classes such as Applying to College and Introduction to Philosophy at juvenile detention centers.

Academic Service

I reviewed for Empirical Methods in Natural Language Processing (EMNLP), 2023; and, I reviewed for the Socially Responsible Language Modelling Research (SoLaR) Workshop at the Conference on Neural Information Processing Systems (NeurIPS), 2023.