The Decoder· Jonathan Kemper·· 2 小时前精选AI 评分51
Sakana AI 与 UC Berkeley 提出 MASS 多智能体自监督方法，可自我迭代但存在共享盲点
Sakana AI's self-improving model gets better at solving and judging, but can't catch its own blind spots
AI 导读
Sakana AI 与 UC Berkeley 提出 MASS 多智能体自监督方法，让同一模型分别承担求解、优化与评估三种角色进行递归自我改进。
来源：The Decoder · the-decoder.com