DimMem: Dimensional Structuring for Efficient Long-Term Agent Memory

TL;DR

DimMem introduces a structured, dimensional memory framework for LLM agents that improves recall accuracy and efficiency by explicitly representing memory units with key attributes, outperforming existing systems.

Contribution

The paper presents DimMem, a novel dimensional memory system that enhances long-term memory in LLM agents through explicit, structured representations, enabling better retrieval and updates.

Findings

DimMem achieves 81.43% and 78.20% accuracy on two benchmarks.

Reduces per-query token cost by 24%.

A compact extractor surpasses larger models after fine-tuning.

Abstract

Large language model (LLM) agents require long-term memory to leverage information from past interactions. However, existing memory systems often face a fidelity--efficiency trade-off: raw dialogue histories are expensive, while flat facts or summaries may discard the structure needed for precise recall. We propose \textbf{DimMem}, a lightweight dimensional memory framework that represents each memory as an atomic, typed, and self-contained unit with explicit fields such as time, location, reason, purpose, and keywords. This representation exposes the structure needed for dimension-aware retrieval, memory update, and selective assistant-context recall without storing full histories in the model context. Across LoCoMo-10 and LongMemEval-S, DimMem achieves \textbf{81.43\%} and \textbf{78.20\%} overall accuracy, respectively, outperforming existing lightweight memory systems while reducing LoCoMo per-query token cost by \textbf{24\%}. We further show that dimensional memory extraction is learnable by compact models: after fine-tuning on the DimMem schema, a Qwen3-4B extractor surpasses LightMem with GPT-4.1-mini on both benchmarks and reaches performance comparable to, or better than, much larger extractors in key settings. These results suggest that explicit dimensional structuring is an effective and efficient foundation for long-term memory in LLM agents. Code is available at https://github.com/ChowRunFa/DimMem.