Epistemic Context Learning: Building Trust the Right Way in LLM-Based Multi-Agent Systems

Ruiwen Zhou, Maojia Song, Xiaobao Wu, Sitao Cheng, Xunjian Yin, Yuxi Xie, Zhuoqun Hao, Wenyue Hua, Liangming Pan, Soujanya Poria, Min-Yen Kan

January 2026

ECL shifts multi-agent reasoning from history-agnostic reference to history-aware trust.

Abstract

Epistemic Context Learning models peer reliability from historical interactions in LLM-based multi-agent systems. The approach conditions predictions on peer profiles, helping agents identify reliable sources and avoid blindly conforming to misleading peers.

Type

Journal article

Publication

arXiv preprint arXiv:2601.21742

Epistemic Context Learning: Building Trust the Right Way in LLM-Based Multi-Agent Systems

Abstract

Ruiwen Zhou

Doctoral Student (Aug ‘25)
Co-Supervised by Soujanya Poria

Yuxi Xie

Doctoral Alumnus (May ‘26). Thesis: Closed-Loop Scaling: Autonomous Improvement of LLM and LVLM Reasoning

Liangming Pan

Doctoral Alumnus (Apr ‘22). Thesis: Towards Generating Deep Questions from Text.

Min-Yen Kan

Associate Professor

Epistemic Context Learning: Building Trust the Right Way in LLM-Based Multi-Agent Systems

Abstract

Ruiwen Zhou

Doctoral Student (Aug ‘25)Co-Supervised by Soujanya Poria

Yuxi Xie

Doctoral Alumnus (May ‘26). Thesis: Closed-Loop Scaling: Autonomous Improvement of LLM and LVLM Reasoning

Liangming Pan

Doctoral Alumnus (Apr ‘22). Thesis: Towards Generating Deep Questions from Text.

Min-Yen Kan

Associate Professor

Doctoral Student (Aug ‘25)
Co-Supervised by Soujanya Poria