EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Self-Healing Memory Architectures for Large Language Model-Based Multi-Agent Collaboration

Toby Walsh, Anton van den Hengel, Stephen J. Roberts

American Journal of Artificial Intelligence and Neural Networks · 01.04.2026 · 7 (1) · 149-159

DOI: 10.71465/ajainn3659 ↗

Kurz gefasst

Fachartikel in der Zeitschrift American Journal of Artificial Intelligence and Neural Networks, online erschienen am 01.04.2026. Verfasst von Toby Walsh und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Multi-agent systems built on large language models (LLMs) frequently exchange contextual memory, creating potential vectors for poisoning through malicious prompts or corrupted state propagation. This study proposes a self-healing memory architecture combining semantic anomaly detection and consensus-based repair. Memory embeddings are continuously monitored using cosine similarity drift detection relative to historical stable clusters. When abnormal divergence exceeds a threshold (δ = 0.35), a cross-agent consensus mechanism reconstructs corrupted segments through majority voting and redundancy checks. Evaluation was conducted on 18 collaborative reasoning tasks involving 10–20 LLM agents. Under simulated adversarial prompt injection, task accuracy declined by 29.5% in baseline settings but only 8.3% under the proposed repair framework. Memory corruption persistence time decreased from 14.2 interaction rounds to 4.7 rounds. The architecture enhances resilience against semantic memory poisoning in collaborative LLM ecosystems.

Zitieren

APA

Walsh, T., Hengel, A. V. D., & Roberts, S. J. (2026). Self-Healing Memory Architectures for Large Language Model-Based Multi-Agent Collaboration. American Journal of Artificial Intelligence and Neural Networks, 7(1), 149-159. https://doi.org/10.71465/ajainn3659

Vancouver

Walsh T, Hengel AVD, Roberts SJ. Self-Healing Memory Architectures for Large Language Model-Based Multi-Agent Collaboration. American Journal of Artificial Intelligence and Neural Networks. 2026;7(1):149-159. doi:10.71465/ajainn3659

BibTeX
@article{Walsh2026,
  author = {Walsh, Toby and Hengel, Anton van den and Roberts, Stephen J.},
  title = {Self-Healing Memory Architectures for Large Language Model-Based Multi-Agent Collaboration},
  journal = {American Journal of Artificial Intelligence and Neural Networks},
  year = {2026},
  volume = {7},
  number = {1},
  pages = {149--159},
  doi = {10.71465/ajainn3659}
}
Kostenlos starten →