微软研究院焦点第97期：智能体框架、编码智能体评测与延伸期天气预报
原标题：Microsoft Research Focus – Edition #97
AI 摘要
微软研究院发布第97期Research Focus通讯，介绍了多项研究进展：自改进框架ForkPilot让AI智能体在需要回溯决策时自适应调整策略，在六个基准测试中达到现有最佳方法水平并将AI处理需求降低最多59%；CABRA框架通过从零生成编码挑战来定位编码智能体的失败原因，发现理解陌生代码库仍是最大障碍；TeleTune从用户活动日志中学习可复用技能与工作流；对Copilot Chat活动的分析显示编辑文档是最常见用途，但各行业在细分任务上差异显著；Shiksha Copilot在印度低资源学校的研究显示其减少了教师备课与行政时间；PBC机器学习框架通过纠正天气模型的历史误差，在改进2至6周延伸期天气预报的全球竞赛中夺冠。
正文节选
What’s new in research? You may already be familiar with the Research Focus newsletter (opens in new tab) on LinkedIn. Now, you can catch up on highlights from across the research community at Microsoft on our social channels, YouTube, and major podcast platforms, too. In this edition of Microsoft Research Focus, host Sowbhagya Suresh shares some of the latest news and cutting-edge research from across Microsoft—from understanding the challenges facing coding agents to uncovering how people use