Temporal Traceability and Source Attribution of Memory Poisoning in Distributed Multi-Agent Systems
Global Media and Social Sciences Research Journal · 25.03.2026 · 7 (1) · 280-286
DOI: 10.71465/gmssrj174 ↗Kurz gefasst
Fachartikel in der Zeitschrift Global Media and Social Sciences Research Journal, online erschienen am 25.03.2026. Verfasst von Lukas Hoffmann und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).
Abstract (Originalsprache)
Identifying the origin of memory poisoning is essential for containment and system recovery. This study develops a temporal traceability model to detect and localize poisoning sources in distributed agent systems. A time-indexed memory influence matrix is constructed to quantify causal propagation strength between agents. Granger causality testing combined with sparse regression identifies probable poisoning origins. Experiments were performed on simulated collaborative planning systems with 80 agents and controlled adversarial injection points. Source attribution accuracy reached 91.2%, and containment strategies reduced secondary contamination by 54.6% within 15 update iterations. The proposed framework improves forensic analysis and containment efficiency in adversarial multi-agent environments.
Zitieren
APA
Hoffmann, L., Becker, J., & Schneider, T. (2026). Temporal Traceability and Source Attribution of Memory Poisoning in Distributed Multi-Agent Systems. Global Media and Social Sciences Research Journal, 7(1), 280-286. https://doi.org/10.71465/gmssrj174
Vancouver
Hoffmann L, Becker J, Schneider T. Temporal Traceability and Source Attribution of Memory Poisoning in Distributed Multi-Agent Systems. Global Media and Social Sciences Research Journal. 2026;7(1):280-286. doi:10.71465/gmssrj174
BibTeX
@article{Hoffmann2026,
author = {Hoffmann, Lukas and Becker, Johannes and Schneider, Tobias},
title = {Temporal Traceability and Source Attribution of Memory Poisoning in Distributed Multi-Agent Systems},
journal = {Global Media and Social Sciences Research Journal},
year = {2026},
volume = {7},
number = {1},
pages = {280--286},
doi = {10.71465/gmssrj174}
}
EUVIMED