MemFold: Learning Compact Soft Memory for Long-Context Personalization via On-Policy Optimization
- 类型：arxiv
- 标识：2609.36435
- 链接：https://arxiv.org/abs/2609.36435
- 主分类：engineering
- 形态：method
- TLDR：An assistant that serves the same user over a long horizon has to answer from what that user has revealed: which preferences still hold, which were revised, and which constraints apply now. Retaining that information is not the same as acting on it, and the two are usually optimized as if they were. Keeping the information as text makes the reader's input grow with the retained history, while compressing it into a fixed number of latent vectors bounds the interface but is typically trained to reconstruct text or imitate reference answers, both of which are scored on sequences the reader never
- 待LLM分类：是
- 来源文件：
- /inbox/tom/_candidates/2026-10-03-agent-rag-longcontext-candidates.json