WFM 把 LLM Wiki 转成图结构来训练智能体记忆
Interesting paper on agent memory stored as a linked markdown wiki. Lots of great ideas and insight...
给做智能体长期记忆的人：WFM 把链接式 markdown wiki 转成图来检索，训练还快了 10.5 倍，方法挺巧。
论文提出 Wiki Foundation Model（WFM），把存储为链接式 markdown 的 LLM Wiki 转成图，用基于查询的消息传递同时利用页面文本和链接结构做检索。团队还设计了 GPU-to-GPU 训练协议，训练速度提升 10.5 倍。论文报告了在五个智能体记忆和多跳推理基准上的成绩。适用场景是长期记忆为链接 markdown 文件夹的智能体。
Interesting paper on agent memory stored as a linked markdown wiki. Lots of great ideas and insight...
Interesting paper on agent memory stored as a linked markdown wiki. Lots of great ideas and insights if you work with LLM Wikis. Wikis are useful for agents because each page holds dense text and the links between pages hold structure. WFM is a Wiki Foundation Model trained to use both at once. It turns an LLM Wiki into a graph and retrieves from it with message passing conditioned on the query, so the text of each page and the link structure shape the result together. The team also built a GPU-to-GPU training protocol that trains 10.5x faster, and reports strong results on five agent memory and multi-hop reasoning benchmarks. If your agent's long-term memory is a folder of linked markdown files, WFM is designed for that format. Paper: arxiv.org/abs/2609.18182 Chat with Paper: academy.dair.ai/papers/wfm-wik… 💬 4 🔄 2 ❤️ 8 👀 1594 📊 7 ⚡