notesum.ai

Published at November 2

Role Play: Learning Adaptive Role-Specific Strategies in Multi-Agent Interactions

cs.MA

cs.AI

cs.LG

Released Date: November 2, 2024

Authors: Weifan Long¹, Wen Wen², Peng Zhai³, Lihua Zhang³

Aff.: ¹Academy for Engineering and Technology, Fudan University; ²School of New Media and Communication, Tianjin University; ³Academy for Engineering and Technology, Fudan University; Ji Hua Laboratory

Arxiv: http://arxiv.org/abs/2411.01166v1

	Harvest			CleanUp
Algorithm	selfish	prosocial	inequity-averse	selfish	prosocial	inequity-averse
AnyPlay	20.62 (4.91)	25.13 (5.29)	3.19 (9.12)	11.58 (8.41)	2.42 (12.16)	-1.90 (4.21)
Trajedi	19.74 (5.21)	23.95 (5.96)	3.07 (11.06)	-3.41(20.87)	-23.00(16.49)	-22.30 (13.05)
BRDiv	12.45 (8.73)	16.46(5.57)	-26.97 (17.13)	8.56 (10.36)	2.03(17.47)	-5.38 (9.41)
HSP	21.71 (8.01)	23.24(14.23)	4.95 (10.95)	10.90 (9.04)	8.15(11.86)	-1.20(3.98)
RP (z= $0$ )	6.94 (12.71)	13.28 (14.06)	6.35 (5.82)	13.23 (11.05)	24.37 (13.45)	8.55 (11.16)
RP (z= $\frac{\pi}{4}$ )	20.46 (3.92)	23.57 (5.56)	15.15 (4.35)	23.35 (13.93)	38.09 (7.71)	23.42 (6.65)
RP (z= $\frac{\pi}{2}$ )	20.60 (3.52)	25.84 (3.98)	15.61 (6.94)	22.73 (12.58))	39.53 (6.99)	18.64 (10.65)