Stephanie Fu
phd student | EECS @ UC Berkeley

I am a PhD student at BAIR advised by Trevor Darrell. I have been supported by the College of Engineering Fellowship and am currently funded by the NSF GRFP.
Previously, I graduated from MIT with an MEng in computer science (advised by Phillip Isola) and bachelors degrees in computer science and music. During my undergrad, I was fortunate to work on exciting research under Bill Freeman, Yoel Fink, and Phillip Isola.
I am broadly interested in computer vision and representation learning. Recently I have been thinking about 1) modeling and understanding human visual intelligence and 2) using that knowledge to build better perceptual representations.
In 2019, I co-founded TEDxMIT and helped launch the inaugural conference at MIT CSAIL. Since then, TEDxMIT has brought communities across the Greater Boston area together with regular events and speakers. Check out past and upcoming conferences here!
news
Aug 2025 | Our extended abstract Incorporating foveal sampling and integration to model 3D shape inferences was selected as an oral at the Conference on Cognitive Computational Neuroscience (CCN) 2025! |
---|---|
Jun 2025 | Our paper Hidden in plain sight: VLMs overlook their visual representations won best paper at the CVPR 2025 EVAL-FoMo 2 Workshop and was accepted to COLM 2025! |
Oct 2024 | We just released When does perceptual alignment benefit vision representations? and will be presenting a poster at NeurIPS 2024! |
selected publications
- CCN oral (top 6.5%)CCN 2025
-
-
- NeurIPS Datasets & Benchmarks 2024
-
- NeurIPS 2023 Spotlight
- NeurIPS 2020 Competition and Demonstration Track
- Nature Communications