RSM-full记忆方法:分离写入合并与读取组装,4k预算达全上下文83%质量

DAIR.AI · @dair_ai · X·2026-09-09 13:20·26分钟前
AI 导读

一项研究为长程智能体提出RSM-full记忆方法,将记忆写入时的合并规则与读取时的内容组装分离。在4k token预算下,该方法达到全上下文质量的83%,仅耗32%的token成本。消融实验显示,合并规则和分组打包器分别贡献5.7点和5.0点提升,并在RealMem基准上超越Budget-RAG、Streaming-Proto和A-MEM基线,与BM25-RAG持平。

DAIR.AI@dair_ai
37AI 编辑部评分,满分 100

RSM-full记忆方法:分离写入合并与读取组装,4k预算达全上下文83%质量

2026-09-09 13:20· 26分钟前
AI 导读

一项研究为长程智能体提出RSM-full记忆方法,将记忆写入时的合并规则与读取时的内容组装分离。在4k token预算下,该方法达到全上下文质量的83%,仅耗32%的token成本。消融实验显示,合并规则和分组打包器分别贡献5.7点和5.0点提升,并在RealMem基准上超越Budget-RAG、Streaming-Proto和A-MEM基线,与BM25-RAG持平。

Good work on improving memory for long-horizon agents.

They separate two things that agent memory papers usually collapse into one. How memories get merged when they are written, and how retrieved content gets assembled into the prompt.

The setting is a tight prompt budget of 2k to 5k tokens, where full-context prompting is off the table because of latency, cost and context limits.

RSM-full combines a cosine-gated max-member merge rule on the write side with an atom-aware grouped packer on the read side. At a 4k budget it reaches 83% of full-context quality at 32% of the token cost.

The ablations attribute the gain to both halves separately. The merge rule is worth 5.7 points over online k-means and matched DP-means. The grouped packer is worth 5.0 points over flat concatenation.

It reproduces on RealMem, beating Budget-RAG, Streaming-Proto and the A-MEM agentic memory baseline, and landing level with BM25-RAG rather than above it.

The authors state that higher-token baselines stay stronger outside this budget range.

Paper: https://academy.dair.ai/papers/compact-memory-llm-agents-via-online-max-member-clustering-and-atom-aware-packin-2609.04915

来源:DAIR.AI· x.com