TY  - RPRT
TI  - Improving Self-Supervised Learning by Characterizing Idealized Representations
AU  - Yann Dubois
AU  - Tatsunori Hashimoto
AU  - Stefano Ermon
AU  - Percy Liang
PY  - 2022
UR  - https://arxiv.org/abs/2209.06235
ID  - 2209.06235
ER  -