BaRe-Mem: Bayesian Reliability Memory for Robust and Adaptive Agent Consultation

BaRe-Mem tracks advisor reliability so multi-agent systems resist misleading consultations better than debate or voting.

BaRe-Mem is an online Bayesian reliability memory for multi-agent consultation. It estimates advisor reliability from a central model's internal beliefs and updates those estimates from historical interactions, then uses them to weight advisor responses and to choose consultation versus autonomous reasoning. Across nine benchmarks and six central models, it is more robust to misleading advisors than debate and majority voting, and on harder tasks it stays above autonomous reasoning at every tested misleading level. Applied to worker allocation, it improves MuSiQue completion over routing by historical success counts and identifies capable workers earlier.