# 加权记忆树：为长时运行智能体引入可恢复记忆

- 来源：DAIR.AI (@dair_ai)
- 发布时间：2026-08-25 05:00
- AIHOT 分数：49
- AIHOT 链接：https://aihot.virxact.com/items/cmt7rkjgz2eizro7325fdho91
- 原文链接：https://x.com/dair_ai/status/2091994046948655506

## AI 摘要

DAIR.AI 介绍一种名为加权记忆树的新方法，为长时运行智能体实现可恢复记忆。它将执行组织为任务、子任务和动作，并为每条记忆赋予动态保留分数，事件更新提升分数、选择衰减降低分数。在 GAIA-Text 上，该方法搭配 Qwen3-8B、Gemma 4 E4B 和 Llama-3.1-8B，平均比线性记忆高出 9.97 分，同时提示词 token 使用量减少 32.8%。

## 正文

Interesting new approach to enable memory in long-running agents.

Weighted Memory Tree organizes execution into tasks, subtasks, and actions, then gives every memory a retention score that moves. Event based updates raise it, selection based decay lowers it.

When a subtask finishes, its step by step detail collapses into a short summary and the full version stays retrievable. If a later step needs the details, the agent pulls them back.

Context trimming is usually permanent. Drop the wrong turn and the agent has no way to recover it. Folding gives you the same token savings with a way back.

On GAIA-Text with Qwen3-8B, Gemma 4 E4B, and Llama-3.1-8B, it beats linear memory by 9.97 points on average while using 32.8% fewer prompt tokens. Memory poisoning experiments show retention scoring limits how far unreliable information spreads.

Paper: https://arxiv.org/abs/2608.20631

Track more trending AI papers in our academy: https://academy.dair.ai/