Memory in the LLM Era: Modular Architectures and Strategies in a Unified Framework

TL;DR

This paper reviews, compares, and analyzes various memory methods in large language model agents, introduces a new superior memory approach, and discusses future research directions.

Contribution

It provides a unified framework for memory methods, a comprehensive experimental comparison, and a novel memory method that outperforms existing techniques.

Findings

Existing memory methods vary in effectiveness across benchmarks.

The new memory method surpasses state-of-the-art performance.

Systematic comparison reveals strengths and weaknesses of different approaches.

Abstract

Memory emerges as the core module in the large language model (LLM)-based agents for long-horizon complex tasks (e.g., multi-turn dialogue, game playing, scientific discovery), where memory can enable knowledge accumulation, iterative reasoning and self-evolution. A number of memory methods have been proposed in the literature. However, these methods have not been systematically and comprehensively compared under the same experimental settings. In this paper, we first summarize a unified framework that incorporates all the existing agent memory methods from a high-level perspective. We then extensively compare representative agent memory methods on two well-known benchmarks and examine the effectiveness of all methods, providing a thorough analysis of those methods. As a byproduct of our experimental analysis, we also design a new memory method by exploiting modules in the existing methods, which outperforms the state-of-the-art methods. Finally, based on these findings, we offer promising future research opportunities. We believe that a deeper understanding of the behavior of existing methods can provide valuable new insights for future research.