RoboMemory: A Brain-inspired Multi-memory Agentic Framework for Interactive Environmental Learning in Physical Embodied Systems
This addresses the problem of memory integration and planning for embodied agents in real-world environments, representing a strong specific gain.
The paper tackled challenges in embodied agents by introducing RoboMemory, a brain-inspired multi-memory framework, which improved average success rates by 25% over its baseline and exceeded the SOTA Gemini-1.5-Pro by 3% on EmbodiedBench.
Embodied agents face persistent challenges in real-world environments, including partial observability, limited spatial reasoning, and high-latency multi-memory integration. We present RoboMemory, a brain-inspired framework that unifies Spatial, Temporal, Episodic, and Semantic memory under a parallelized architecture for efficient long-horizon planning and interactive environmental learning. A dynamic spatial knowledge graph (KG) ensures scalable and consistent memory updates, while a closed-loop planner with a critic module supports adaptive decision-making in dynamic settings. Experiments on EmbodiedBench show that RoboMemory, built on Qwen2.5-VL-72B-Ins, improves average success rates by 25% over its baseline and exceeds the closed-source state-of-the-art (SOTA) Gemini-1.5-Pro by 3%. Real-world trials further confirm its capacity for cumulative learning, with performance improving across repeated tasks. These results highlight RoboMemory as a scalable foundation for memory-augmented embodied intelligence, bridging the gap between cognitive neuroscience and robotic autonomy.