notesum.ai

Published at October 31

Rethinking Inverse Reinforcement Learning: from Data Alignment to Task Alignment

cs.LG

cs.AI

Released Date: October 31, 2024

Authors: Weichao Zhou¹, Wenchao Li¹

Aff.: ¹Boston University

Arxiv: http://arxiv.org/abs/2410.23680v1

Refer to caption

	RECOIL	PAGAR-RECOIL
hopper-random	$106.87\pm 2.69$	$111.16\pm 0.51$
halfcheetah-random	$80.84\pm 17.62$	$92.94\pm 0.10$
walker2d-random	$108.40\pm 0.04$	$108.40\pm 0.12$
ant-random	$113.34\pm 2.78$	$121\pm 5.86$