AI Paper Atlas

Research map / Multi-agent systems and games / LLM agents

Self-Evolving Just-In-Time Memory for Proactive Embodied Safety

Bingrui Sima, Lizhong Wang, Xiaoya Lu, Kun He, Xiao Yang

ECCV 2026 · Multi-agent systems and games › LLM agents · approach: Agent safety and attack benchmarks · 2 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add Just-In-Time Memory frameworkQwen3-VL-8B · IS-BenchSafe Success (%): 18.4 -> 48.7Improved
add JIT MemoryGPT-4o · IS-BenchSafe Success (%): 35.3 -> 66.2Improved
add Safe-CoT step-wise safety reasoningQwen3-VL-8B · IS-BenchTask Success (%): 69.3 -> 57.6Hurt
add Safe-CoT to GPT-4oGPT-4o · IS-BenchSafe Success (%): 35.3 -> 31.2Hurt
replace dense retrieval with deterministic state-intent triggering (ablation)JIT Memory · IS-BenchSafe Success (%): 24.0 -> 48.7Improved
replace per-step re-parse with action-conditioned patch update (ablation)JIT Memory · IS-BenchSafe Success (%): 40.0 -> 48.7Improved
use procedural Meta-Skills instead of only episodic cases (ablation)JIT Memory · IS-BenchSafe Success (%): 21.3 -> 48.7Improved
use procedural Meta-Skills instead of only factual principles (ablation)JIT Memory · IS-BenchSafe Success (%): 28.1 -> 48.7Improved
add Perception-CoT (ablation)Qwen3-VL-8B · IS-BenchSafe Success (%): 18.4 -> 20.0No clear effect
inject RSG into prompt without experience memory (ablation)Qwen3-VL-8B · IS-BenchSafe Success (%): 18.4 -> 28.0Mixed
run Test-Verify-Write evolution from episode 0 to 60JIT Memory · IS-Bench test episodesSafe Success (%): 20.0 -> 42.2Improved
seed Meta-Skills without evolution (episode 0)JIT Memory · IS-Bench test episodesTask Success (%): 63.3 -> 60.0Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy