Robust Memory Update Mechanisms Against Poisoning Attacks in Multi-Agent Reinforcement Learning
Frontiers in Artificial Intelligence Research · 15.03.2026 · 3 (2) · 24-32
DOI: 10.71465/fair756 ↗Kurz gefasst
Fachartikel in der Zeitschrift Frontiers in Artificial Intelligence Research, online erschienen am 15.03.2026. Verfasst von James Thompson und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
In multi-agent reinforcement learning (MARL), shared replay buffers and inter-agent memory exchange introduce vulnerability to memory poisoning attacks. This work proposes a confidence-weighted memory validation mechanism integrated into the experience sharing pipeline. Each memory entry is assigned a trust score derived from temporal consistency and reward deviation metrics. A Bayesian filtering process excludes anomalous transitions before propagation to peer agents. Experiments were conducted on cooperative navigation and resource allocation benchmarks with 12–24 agents. Under a 20% poisoning injection rate, baseline MARL performance dropped by 37.8%, whereas the proposed defense limited degradation to 11.4%. Convergence time improved by 23.6% compared to anomaly-blind training.The method effectively mitigates adversarial memory contamination in collaborative reinforcement learning environments.
Zitieren
APA
Thompson, J., Bennett, O., & Carter, D. (2026). Robust Memory Update Mechanisms Against Poisoning Attacks in Multi-Agent Reinforcement Learning. Frontiers in Artificial Intelligence Research, 3(2), 24-32. https://doi.org/10.71465/fair756
Vancouver
Thompson J, Bennett O, Carter D. Robust Memory Update Mechanisms Against Poisoning Attacks in Multi-Agent Reinforcement Learning. Frontiers in Artificial Intelligence Research. 2026;3(2):24-32. doi:10.71465/fair756
BibTeX
@article{Thompson2026,
author = {Thompson, James and Bennett, Oliver and Carter, Daniel},
title = {Robust Memory Update Mechanisms Against Poisoning Attacks in Multi-Agent Reinforcement Learning},
journal = {Frontiers in Artificial Intelligence Research},
year = {2026},
volume = {3},
number = {2},
pages = {24--32},
doi = {10.71465/fair756}
}
EUVIMED