BMAM: Brain-inspired Multi-Agent Memory Framework

2026-01-28

AI Summary

This work addresses the degradation of temporal consistency and behavioral coherence in language model agents during long-term interactions—a phenomenon termed “soul erosion”—by proposing a brain-inspired multi-agent memory framework. Drawing on cognitive science, the framework introduces a multi-system memory architecture into multi-agent systems for the first time, explicitly constructing an episodic memory timeline and integrating three functional subsystems: semantic, salience-aware, and control-oriented, to enable collaborative long-horizon reasoning. Key technical contributions include timeline-based modeling of episodic memory, multi-signal fused retrieval, and functional decomposition of memory processes. Evaluated on the LoCoMo benchmark, the approach achieves 78.45% accuracy, with ablation studies confirming the critical role of hippocampus-inspired episodic memory in temporal reasoning.

Technology Category

• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Multiagent Systems: Multiagent Learning
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search
• User Modeling, Personalization and Recommendation: Studies of user behavior, including longitudinal effects of personalized systems

Abstract

Language-model-based agents operating over extended interaction horizons face persistent challenges in preserving temporally grounded information and maintaining behavioral consistency across sessions, a failure mode we term soul erosion. We present BMAM (Brain-inspired Multi-Agent Memory), a general-purpose memory architecture that models agent memory as a set of functionally specialized subsystems rather than a single unstructured store. Inspired by cognitive memory systems, BMAM decomposes memory into episodic, semantic, salience-aware, and control-oriented components that operate at complementary time scales. To support long-horizon reasoning, BMAM organizes episodic memories along explicit timelines and retrieves evidence by fusing multiple complementary signals. Experiments on the LoCoMo benchmark show that BMAM achieves 78.45 percent accuracy under the standard long-horizon evaluation setting, and ablation analyses confirm that the hippocampus-inspired episodic memory subsystem plays a critical role in temporal reasoning.

Problem

• soul erosion
• long-horizon interaction
• temporal consistency
• behavioral consistency
• memory preservation

Innovation

• brain-inspired memory
• multi-agent memory
• episodic memory
• long-horizon reasoning
• functional memory decomposition