Scholar
Druv Pai
Google Scholar ID: GdOmgYwAAAAJ
PhD Student, UC Berkeley
deep learning theory
computer vision
NLP
Follow
Homepage
↗
Google Scholar
↗
Citations & Impact
All-time
Citations
319
H-index
10
i10-index
10
Publications
17
Co-authors
10
list available
Contact
Email
druvpai@berkeley.edu
CV
Open ↗
Twitter
Open ↗
GitHub
Open ↗
Publications
6 items
On the Edge of Memorization in Diffusion Models
2025
Cited
0
Attention-Only Transformers via Unrolled Subspace Denoising
2025
Cited
1
Simplifying DINO via Coding Rate Regularization
2025
Cited
0
A Global Geometric Analysis of Maximal Coding Rate Reduction
International Conference on Machine Learning · 2024
Cited
10
Scaling White-Box Transformers for Vision
arXiv.org · 2024
Cited
3
Independent and Decentralized Learning in Markov Potential Games
arXiv.org · 2022
Cited
18
Resume (English only)
Academic Achievements
Learning Deep Representations of Data Distributions – Open-source textbook.
Token Statistics Transformer: Linear-Time Attention via Variational Rate Reduction – ICLR 2025 (Spotlight).
On the Edge of Memorization in Diffusion Models.
White-Box Transformers via Sparse Rate Reduction: Compression is all There Is? – JMLR 2024 (parts presented at NeurIPS 2023, ICLR 2024, CPAL 2024).
Simplifying DINO via Coding Rate Regularization – ICML 2025.
Active-Dormant Attention Heads: Mechanistically Demystifying Extreme-Token Phenomena in LLMs.
Co-authors
10 total
Yi Ma (马毅)
Director of School of Computing & Data Science, HKU; Visiting Professor of EECS, Berkeley
Yaodong Yu
University of California, Berkeley
Sam Buchanan
UC Berkeley
Shengbang Tong
NYU Courant
Tianzhe Chu
The University of Hong Kong
Manxi Wu
University of California, Berkeley
Shankar Sastry
University of California
Chinmay Maheshwari
Assistant Professor, Johns Hopkins University
×
Welcome back
Sign in to Agora
Welcome back! Please sign in to continue.
Email address
Password
Forgot password?
Continue
Do not have an account?
Sign up