Salesforce AI Research 提出即时记忆方法，智能体存原始轨迹而非摘要
Salesforce 这篇论文很实用：智能体别急着存摘要，先存原始轨迹，来新任务再提取，三个基准上多涨十几分。
Salesforce AI Research 发布一篇关于智能体记忆的论文，提出 Just-in-Time Memory 方法。核心做法是保存原始任务轨迹，等新任务到来时再决定从中提取什么，而不是在任务结束时生成摘要。一个 curator 模块读取检索到的轨迹和新任务，写出针对该任务的简短记忆载荷，因此可以用该任务的成功与否来训练 curator。在 ALFWorld、WebShop 和 tau2-bench 三个基准上，该方法分别比最强基线高 16.2、16.3 和 3.9 个成功率点，连未训练的 curator 也能追平或超过任务结束时写入的记忆。
Banger paper from Salesforce AI Research on agent memory.
The overall finding is that you want to store raw trajectories and decide what to extract from them when the next task arrives, instead of summarizing each run when it ends.
Just-in-Time Memory uses a curator that reads the retrieved traces together with the new task and writes a short memory payload for that task. Because the payload is used right away, the curator can be trained on whether that same task succeeds.
On ALFWorld, WebShop and tau2-bench it beats the strongest baseline by 16.2, 16.3 and 3.9 success-rate points. Even the untrained curator matches or beats memory that is written when a task ends.
Paper: https://t.co/0TG98s24dt
