杜克、Meta与普林斯顿推出 PAHF 框架
burkov · x · 2026-08-14
A joint research effort from Duke University, Meta, and Princeton introduces PAHF, a novel framework designed to help AI agents adapt to evolving user preferences.
By leveraging live interaction and explicit memory, PAHF enables agents to continually learn individual preferences. The paper notes that this dynamic approach significantly outperforms existing static-data methods.
「编程与Agent」频道最新
- WorldBox 沙盒让 GPT-6 Astra 在 MazeBench 上得分翻倍，从 23 颗宝石涨到 61 颗 — Lianhuiq · 2026-10-16
- CUA 团队回应：Hyprland 0.56.2-4 已支持，Omarchy 升级后即可用 — Teknium · 2026-10-16
- 用 Void 加 fate 一天搭出完整 vibe coding 平台并部署到 Cloudflare — cnakazawa · 2026-10-16
- 一人公司新玩法：8个专职Agent分道编码，架构师本人只做审批 — bradneuberg · 2026-10-16
- 一人带 8 个专家 Agent 干活：两个月搭出工作流，两周抵一年工作量 — MiddleBrownie · 2026-10-16
- 小老板给 AI 助手 Dot 配了独立手机号,替他回客户消息 — Early-Protection2386 · 2026-10-16