【新智元导读】刚刚，由SciMaster团队推出的AI机器学习专家ML-Master 2.0，基于国产开源大模型DeepSeek，在OpenAI权威基准测试MLE-bench中一举击败Google、Meta、微软等国际顶流，刷新全球SOTA，再次登顶 ...

这篇综述揭示了果蝇MLE（DHX9同源物）解旋酶通过结合新型增强子663和双启动子，调控核受体FTZ-F1（NR5A3）组成型（ftz-f1-B）和 ...

OpenAI 今天发布了一个名为 MLE-bench 的基准测试，专门用来测试 AI Agent 的机器学习工程能力！这是要让 AI 自己训练模型、准备数据集、跑实验的节奏吗？！🤯 MLE-bench 是什么？ MLE-bench 是一个离线的 Kaggle 竞赛（机器学习比赛）环境，包含 75 个来自 Kaggle 的机器 ...

【新智元导读】刚刚，由上海交通大学人工智能学院Agents团队提出的AI专家智能体，在OpenAI权威基准测试MLE-bench中击败了业界AI顶流微软，夺冠登顶！ 就在刚刚，一支来自中国高校的团队成功刷榜了OpenAI发布的权威基准测试MLE-bench！ 这一次，荣耀属于上海交通 ...

Want smarter insights in your inbox? Sign up for our weekly newsletters to get only what matters to enterprise AI, data, and security leaders. Subscribe Now OpenAI has introduced a new tool to measure ...

NBA stars are breaking the bank this offseason. Whether it's $270 million for Scottie Barnes, $212 million for O.G. Anunoby or $166 million for Bam Adebayo, players know their worth and are taking ...

OpenAI scientists have designed MLE-bench — a compilation of 75 extremely difficult tests that can assess whether a future advanced AI agent is capable of modifying its own code and improving itself.