Taming “Zombie” Agents: A Markov State-Aware Framework for Resilient Multi-Agent Evolution

Taolin Zhang, Pukun Zhao, Qizhou Chen, Jiuheng Wan, Chen Chen, Xiaofeng He, Chengyu Wang, Richang Hong · 2026

Read Taming “Zombie” Agents: A Markov State-Aware Framework for Resilient Multi-Agent Evolution by Taolin Zhang, Pukun Zhao, Qizhou Chen, Jiuheng Wan, Chen Chen, Xiaofeng He, Chengyu Wang, Richang Hong (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-17348.txt.

Description

Zhang and colleagues introduce AgentRevive, a framework that stops multi-agent systems from permanently discarding agents that could still recover. Instead of hard pruning, agents move among Active, Standby and Terminated states under a risk-aware policy, with standby agents kept for later rounds. Edges follow learned states, so only terminated nodes leave for good. Tests show it beats strong baselines and cuts token use through state-aware scheduling.

Catalog record

CREATORTaolin Zhang, Pukun Zhao, Qizhou Chen, Jiuheng Wan, Chen Chen, Xiaofeng He, Chengyu Wang, Richang Hong
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE62029
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.17348

Browse the full library · About the Tower of Babel