Surya Ganguli
Surya Ganguli is an associate professor of applied physics at Stanford, where his research covers how both biological and artificial neural networks learn. He triple majors at MIT, completes a string theory PhD at UC Berkeley in 2004, and works as a postdoc at UC San Francisco before joining Stanford. His 2015 paper with Jascha Sohl-Dickstein, Eric Weiss and Niru Maheswaranathan describes a generative model that learns to reverse a diffusion process. He co-authors papers that receive an ICML 2021 Outstanding Paper Honorable Mention and a NeurIPS 2022 Outstanding Paper. In 2025 he receives a senior fellowship in the Schmidt Sciences AI2050 program.
- Now at
- Stanford University
- Works
- 275
Publication numbers from OpenAlex, as of 14 October 2026. Citations and h-index are hidden because this OpenAlex record may be incomplete or mix in other authors.
Timeline
-
1998–2004
A string theory PhD
He triple majors at MIT in 1998, then completes a string theory PhD at UC Berkeley in 2004.
-
By 2012
From UCSF postdoc to Stanford
He works as a postdoc in theoretical neuroscience at UCSF, and by 2012 he is an assistant professor at Stanford.
-
2015
A reverse-diffusion generative model
In 2015 he co-writes a paper with Jascha Sohl-Dickstein, Eric Weiss and Niru Maheswaranathan, describing a generative model that learns to reverse a diffusion process.
-
2017
A visiting role at Google
In 2017 he takes a visiting research professorship at Google.
-
2021
Self-supervised learning without pairs
He is one of three authors, with Yuandong Tian and Xinlei Chen, of Understanding self-supervised learning dynamics without contrastive pairs, an ICML 2021 Outstanding Paper Honorable Mention.
-
2022
Beating power-law scaling
He is one of five authors of Beyond neural scaling laws: beating power law scaling via data pruning, a NeurIPS 2022 Outstanding Paper.
-
2025
A Schmidt AI2050 fellowship
In 2025 he receives a senior fellowship in the Schmidt Sciences AI2050 program.
-
As of October 2026
Associate professor at Stanford
He is an associate professor of applied physics at Stanford and a senior fellow at the Stanford Institute for Human-Centered AI. His research spans neuroscience, machine learning and physics, focusing on how both biological and artificial neural networks learn.
Sources and credits
Photo: ganguli-gang.stanford.edu; Neural Dynamics and Computation Lab; creator not stated. License: Not stated by source. Source: ganguli-gang.stanford.edu. Cropped to a square and resized. Ask us to remove or replace this photo.
- Surya Ganguli
- Surya Ganguli | Stanford HAI
- Surya Ganguli - Wikipedia
- Deep Unsupervised Learning using Nonequilibrium Thermodynamics
- [1503.03585] Deep Unsupervised Learning using Nonequilibrium Thermodynamics
- ICML Poster Understanding self-supervised learning dynamics without contrastive pairs
- NeurIPS Poster Beyond neural scaling laws: beating power law scaling via data pruning
- Surya Ganguli named AI2050 Senior Fellow | Stanford School of Humanities and Sciences
- Surya Ganguli | Computer Science