AI Daybook
AI Daybook
오늘의 지면
논문에이전트

MemoryDojo, LLM 에이전트 메모리 중독 공격 벤치마크

MemoryDojo: Benchmarking Poisoning Through LLM Agent Memory

매체
Takara TLDR
게시
중요도
2/5
화제 지수
—

한국어 요약

LLM 에이전트 메모리 시스템이 악의적 지시를 저장했다가 나중 세션에서 회수해 실행하는 '중독' 공격에 노출되는 문제를 다룬다. 155개 멀티 세션 궤적, 4개 도메인, 5개 메모리 시스템을 포함한 벤치마크 MemoryDojo를 제안했다. 모든 메모리 시스템이 중독에 취약하며 그 영향이 아키텍처마다 크게 달라짐을 보였고, 중독 효율을 높이는 메모리 인식 최적화 방법도 제시했다.

원문 발췌 · English

LLM agents benefit from using memory systems to improve performance on long-horizon, multi-session tasks. However, memory introduces a critical attack surface. Malicious instructions encountered in one session may be stored and propagated through memory before being retrieved and acted upon in later sessions. Because memory systems differ in how they process and retrieve information, the downstream effects of poisoning can vary significantly across systems. To quantify the effects of poisoning across different memory architectures, we introduce MemoryDojo, a benchmark for tracing the lifecycle of poisoned information through agent memory and measuring its influence on subsequent agent behavior. MemoryDojo is easily extensible, consists of 155 multi-session long-horizon trajectories covering four domains, and supports five different memory systems. Through our evaluation, we find that all memory systems are susceptible to memory poisoning, with substantial variation in downstream effects. We further propose a memory-aware poison optimization process that increases the poisoning efficacy, highlighting differences between memory architectures. We hope that MemoryDojo will support research on understanding and mitigating agentic memory risks.

#benchmark#memory#agent security#poisoning
원문 기사 읽기
전체 보기 →