notesum.ai

Published at November 8

Game-theoretic LLM: Agent Workflow for Negotiation Games

cs.AI

cs.CL

cs.GT

cs.LG

cs.MA

Released Date: November 8, 2024

Authors: Wenyue Hua¹, Ollie Liu², Lingyao Li³, Alfonso Amayuelas⁴, Julie Chen⁵, Lucas Jiang⁵, Mingyu Jin¹, Lizhou Fan⁶, Fei Sun⁷, William Wang⁴, Xintong Wang¹, Yongfeng Zhang¹

Aff.: ¹Rutgers University; ²University of Southern California; ³University of South Florida; ⁴University of California, Santa Barbara; ⁵Independent Researcher; ⁶Harvard University; ⁷Institute of Computing Technology

Arxiv: http://arxiv.org/abs/2411.05990v2

$\text{Difficulty}(d)$	-2	-3	-4	-5	-6	-7	-8	-9	-10	-11
total number of datapoints	13	27	57	85	108	133	177	189	210	217
Agreement rate	0.5385	0.5556	0.5614	0.6235	0.6574	0.6917	0.7119	0.7249	0.7381	0.7373
envy free rate	0.3077	0.4074	0.4035	0.4824	0.5463	0.6015	0.6441	0.6614	0.6810	0.6820
Pareto optimal rate	0.5384	0.4444	0.4385	0.4823	0.5277	0.5413	0.5310	0.5396	0.5523	0.5529
envy free and Pareto optimal rate	0.3077	0.3333	0.3333	0.3882	0.4537	0.4812	0.4858	0.4973	0.5142	0.5161