Google给AI装了个'错误百科全书',让AI能从失败中学习,小模型也能有大模型的表现。
Google Research推出WikiSkill框架,让AI代理能够记录成功和失败经验。该框架采用类似维基的结构保存知识,使AI代理在每次运行后不会丢弃学习内容。研究显示,大型模型从WikiSkill中获益更多,而小型模型应用该框架后可达到未使用框架的大型模型的性能水平。
Google's WikiSkill gives AI agents a persistent memory of past mistakes to sharpen future performance
Google Research has introduced WikiSkill, a framework that gives AI agents a persistent knowledge base. Instead of discarding what they learned after each run, agents document both failures and successes in a wiki-like structure and use that knowledge to get better over time. Larger models benefit more, but smaller models with WikiSkill can match the performance of larger ones without it. The article Google's WikiSkill gives AI agents a persistent memory of past mistakes to sharpen future performance appeared first on The Decoder .
- arXiv cs.AI08-27 17:59原文