notesum.ai

Published at November 27

Scalable Multi-Objective Reinforcement Learning with Fairness Guarantees using Lorenz Dominance

cs.LG

Released Date: November 27, 2024

Authors: Dimitris Michailidis¹, Willem Röpke², Diederik M. Roijers³, Sennay Ghebreab², Fernando P. Santos¹

Aff.: ¹Socially-Intelligent Artificial Systems, Informatics Institute, University of Amsterdam; ²AI Lab, Vrije Universiteit Brussel; ³AI Lab, Vrije Universiteit Brussel & Innovation Department, City of Amsterdam

Arxiv: http://arxiv.org/abs/2411.18195v1

	GPI-LS	PCN	LCN
HV	$22622.8\pm 55.4$	$22845.4\pm 9.6$	$22838.0\pm 0.0$
EUM	$53.86\pm 0.02$	$53.82\pm 0.02$	$53.76\pm 0.03$