R.I.P.
π»
Ghosted
PropUQ-MAS: Propagation-Aware Uncertainty Quantification for LLM Multi-Agent Systems
August 22, 2026 Β· Grace Period Β· π EMNLP 2026
Authors
Yaokun Liu, Yifan Liu, Daniel Yue Zhang, Ruichen Yao, Zelin Li, Dong Wang
arXiv ID
2608.22130
Category
cs.MA: Multiagent Systems
Cross-listed
cs.CL
Citations
0
Venue
EMNLP 2026
Abstract
LLM-based multi-agent systems (MAS) solve complex tasks through communication among role-specialized agents. However, inter-agent dependencies introduce reliability risks beyond isolated agent failures. For instance, errors in intermediate messages could be inherited and amplified by downstream agents. Existing uncertainty quantification (UQ) methods mainly target isolated responses or single-agent reasoning, and therefore fail to capture uncertainty propagation in MAS. To this end, we propose PropUQ-MAS, an error propagation-aware UQ framework that represents MAS execution as a communication-structured graph and estimates each step's reliability by combining local uncertainty with uncertainty inherited from upstream messages. Extensive experiments demonstrate that PropUQ-MAS consistently improves UQ in MAS, with average relative gains of +6.10% in AUROC and +47.58% in PRR.
Community Contributions
Found the code? Know the venue? Think something is wrong? Let us know!
π Similar Papers
In the same crypt β Multiagent Systems
R.I.P.
π»
Ghosted
Mean Field Multi-Agent Reinforcement Learning
π
π
The Cartographer
A Survey and Critique of Multiagent Deep Reinforcement Learning
π
π
The Cartographer
A Survey of Learning in Multiagent Environments: Dealing with Non-Stationarity
π
π
The Cartographer
Collaborative vehicle routing: a survey
R.I.P.
π»
Ghosted