[人人能懂] AI的内功心法：少即是多，巧胜于拙

00:00:38 给AI的大脑做减法：一种聪明的“偷懒”智慧00:04:34 你的数据正在“减肥”：人工智能时代的生存新法则00:09:23 AI的“闭关修炼”：不给它网络，它能有多强？00:13:07 给AI装上一个“语法”导航仪00:16:49 AI的“通感”：为什么“我写的”和“我是作者”是一回事？本期介绍的几篇论文：[LG] XQuant: Breaking the Memory Wall for LLM Inference with KV Cache Rematerialization  [UC Berkeley & FuriosaAI]  https://arxiv.org/abs/2508.10395---[LG] SoK: Data Minimization in Machine Learning  [ETH Zurich]  https://arxiv.org/abs/2508.10836---[CL] SSRL: Self-Search Reinforcement Learning  [Tsinghua University & Shanghai AI Laboratory]  https://arxiv.org/abs/2508.10874---[LG] Constrained Decoding of Diffusion LLMs with Context-Free Grammars  [ETH Zurich]  https://arxiv.org/abs/2508.10111---[CL] A Rose by Any Other Name Would Smell as Sweet: Categorical Homotopy Theory for Large Language Models  [Adobe Research]  https://arxiv.org/abs/2508.10018在小宇宙查看该单集文稿