About

I am a postdoctoral researcher in the Department of Computer Science at Princeton University, where I work with Tom Griffiths. I received my Ph.D. from the Department of Computer Science at the University of Toronto and the Vector Institute, where my advisor was Richard Zemel. Prior to my Ph.D., I received my M.Sc. in Computer Science from the same department and my B.Sc. in Biomedical Engineering from Yale University.

I am broadly interested in the intersection between deep learning and Bayesian inference. In my research, I use insights from Bayesian inference to design deep neural networks that are resilient and reliable, even in new environments.

Research interests: metalearning, hierarchical Bayesian models, continual learning, Bayesian filtering, nonparametric statistics, uncertainty quantification, distribution-free inference.

I am on the academic job market for 2024-2025!

News

Sep. 2024: One paper accepted at NeurIPS 2024.
Jul. 2024: One paper accepted at TMLR.
Feb. 2024: I gave an invited talk at Stanford University in the Department of Statistics.
Jan. 2024: Two papers accepted at ICLR 2024.
Sep. 2023: Two papers accepted at NeurIPS 2023.

Selected Publications

A Metalearned Neural Circuit for Nonparametric Bayesian Inference

Jake C. Snell, Gianluca Bencomo, Thomas L. Griffiths

To appear in: Neural Information Processing Systems (NeurIPS) 2024.
arXiv

Implicit Maximum a Posteriori Filtering via Adaptive Optimization

Gianluca M. Bencomo, Jake C. Snell, Thomas L. Griffiths

International Conference on Learning Representations (ICLR) 2024.
PDF

Quantile Risk Control: A Flexible Framework for Bounding the Probability of High-Loss Predictions

Jake C. Snell, Thomas P. Zollo, Zhun Deng, Toniann Pitassi, Richard Zemel

International Conference on Learning Representations (ICLR) 2023.
PDF

Bayesian Few-Shot Classification with One-vs-Each PĆ³lya-Gamma Augmented Gaussian Processes

Jake C. Snell, Richard Zemel

International Conference on Learning Representations (ICLR) 2021.
PDF
arXiv
Code

Prototypical Networks for Few-shot Learning

Jake C. Snell, Kevin Swersky, Richard Zemel

Neural Information Processing Systems (NeurIPS) 2017.