Edward Hu is building his own company. He was previously a researcher at OpenAI and received his research training as a Ph.D. student advised by Yoshua Bengio, a recipient of the 2018 A.M. Turing Award.

Before graduate school, Edward was a researcher at Microsoft, where he invented LoRA and μTransfer. LoRA is now one of the most popular methods for customizing AI models, and μTransfer is underpinning the largest AI models being developed today.

Selected Publications

Amortizing Intractable Inference in Large Language Models
ICLR 2024 (oral presentation)
[Paper | Code]

GFlowNet-EM for Learning Compositional Latent Variable Models
ICML 2023
[Paper | Code]

μTransfer: Tuning Large Neural Networks via Zero-Shot Hyperparameter Transfer
NeurIPS 2021
[Paper | Blog | Code | Video | TheRegister | TechRadar | AIM]

LoRA: Low-Rank Adaptation of Large Language Models
ICLR 2022
[Paper | Video | Code | Mentioned in "The innovation behind AI at Scale"]

GFlowNet Foundations
JMLR 2023
[Paper]

μP: Feature Learning in Infinite-Width Neural Networks
ICML 2021
[Paper | Blog | Code]

Improved Image Wasserstein Attacks and Defenses
Best Paper at ICLR Trustworthy ML Workshop 2020
[Paper | Code]

Randomized smoothing of all shapes and sizes
ICML 2020
[Paper | Code]

This website is based on Indigo by Kopplin.