BrainMem: Brain-Inspired Evolving Memory for Embodied Agent Task Planning

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Ma, Xiaoyu, Hu, Lianyu, Tang, Wenbing, Hu, Zixuan, Liao, Zeqin, Wu, Zhizhen, Liu, Yang
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866918452949680128
author Ma, Xiaoyu
Hu, Lianyu
Tang, Wenbing
Hu, Zixuan
Liao, Zeqin
Wu, Zhizhen
Liu, Yang
author_facet Ma, Xiaoyu
Hu, Lianyu
Tang, Wenbing
Hu, Zixuan
Liao, Zeqin
Wu, Zhizhen
Liu, Yang
contents Embodied task planning requires agents to execute long-horizon, goal-directed actions in complex 3D environments, where success depends on both immediate perception and accumulated experience across tasks. However, most existing LLM-based planners are stateless and reactive, operating without persistent memory and therefore repeating errors and struggling with spatial or temporal dependencies. We propose BrainMem(Brain-Inspired Evolving Memory), a training-free hierarchical memory system that equips embodied agents with working, episodic, and semantic memory inspired by human cognition. BrainMem continuously transforms interaction histories into structured knowledge graphs and distilled symbolic guidelines, enabling planners to retrieve, reason over, and adapt behaviors from past experience without any model fine-tuning or additional training. This plug-and-play design integrates seamlessly with arbitrary multi-modal LLMs and greatly reduces reliance on task-specific prompt engineering. Extensive experiments on four representative benchmarks, including EB-ALFRED, EB-Navigation, EB-Manipulation, and EB-Habitat, demonstrate that BrainMem significantly enhances task success rates across diverse models and difficulty subsets, with the largest gains observed on long-horizon and spatially complex tasks. These results highlight evolving memory as a promising and scalable mechanism for generalizable embodied intelligence.
format Preprint
id arxiv_https___arxiv_org_abs_2604_16331
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle BrainMem: Brain-Inspired Evolving Memory for Embodied Agent Task Planning
Ma, Xiaoyu
Hu, Lianyu
Tang, Wenbing
Hu, Zixuan
Liao, Zeqin
Wu, Zhizhen
Liu, Yang
Robotics
Artificial Intelligence
Computer Vision and Pattern Recognition
Multiagent Systems
Embodied task planning requires agents to execute long-horizon, goal-directed actions in complex 3D environments, where success depends on both immediate perception and accumulated experience across tasks. However, most existing LLM-based planners are stateless and reactive, operating without persistent memory and therefore repeating errors and struggling with spatial or temporal dependencies. We propose BrainMem(Brain-Inspired Evolving Memory), a training-free hierarchical memory system that equips embodied agents with working, episodic, and semantic memory inspired by human cognition. BrainMem continuously transforms interaction histories into structured knowledge graphs and distilled symbolic guidelines, enabling planners to retrieve, reason over, and adapt behaviors from past experience without any model fine-tuning or additional training. This plug-and-play design integrates seamlessly with arbitrary multi-modal LLMs and greatly reduces reliance on task-specific prompt engineering. Extensive experiments on four representative benchmarks, including EB-ALFRED, EB-Navigation, EB-Manipulation, and EB-Habitat, demonstrate that BrainMem significantly enhances task success rates across diverse models and difficulty subsets, with the largest gains observed on long-horizon and spatially complex tasks. These results highlight evolving memory as a promising and scalable mechanism for generalizable embodied intelligence.
title BrainMem: Brain-Inspired Evolving Memory for Embodied Agent Task Planning
topic Robotics
Artificial Intelligence
Computer Vision and Pattern Recognition
Multiagent Systems
url https://arxiv.org/abs/2604.16331