Embedded Files
I'm most interested in studying symmetries in Deep Learning. Specifically, I mostly work on neural parameter symmetries, which are operators which can be applied to neural network weights without changing the function they represent.
See a talk I gave on Neural Parameter Symmetries here (pwd: Sgw&z8fZ)
* represents equal contribution
Papers
The Empirical Impact of Neural Parameter Symmetries, or Lack Thereof
The Empirical Impact of Neural Parameter Symmetries, or Lack Thereof
Derek Lim*, Theo (Moe) Putterman*, Robin Walters, Haggai Maron, Stefanie Jegelka
Accepted to NeurIPS 2024, Won Best Paper Award at HiLD ICML 2024
Learning on LoRAs:
Learning on LoRAs:
GL-EQUIVARIANT PROCESSING OF LOW-RANK
WEIGHT SPACES FOR LARGE FINETUNED MODELS
Theo (Moe) Putterman*, Derek Lim*, Yoav Gelberg, Haggai Maron, Stefanie Jegelka
Submitted to ICML 2025
Page updated
Report abuse