The Perils of Optimizing Learned Reward Functions: Low Training Error Does Not Guarantee Low Regret

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Fluri, Lukas, Lang, Leon, Abate, Alessandro, Forré, Patrick, Krueger, David, Skalse, Joar
Format:	Preprint
Published:	2024
Subjects:	Machine Learning Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Internet

https://arxiv.org/abs/2406.15753