TY  - RPRT
TI  - A General Theoretical Paradigm to Understand Learning from Human Preferences
AU  - Mohammad Gheshlaghi Azar
AU  - Mark Rowland
AU  - Bilal Piot
AU  - Daniel Guo
AU  - Daniele Calandriello
AU  - Michal Valko
AU  - Rémi Munos
PY  - 2023
UR  - https://arxiv.org/abs/2310.12036
ID  - 2310.12036
ER  -