你有没有想过,AI的“脑子”里到底在想些什么?这一期,我们就来当一回“AI心理学家”,从几篇最新论文出发,探寻AI的内心世界:看它如何“自己教自己”实现顿悟,又为何会陷入“学不动”的瓶颈;我们会揭秘它那张写满内心独白的“草稿纸”,看看它是否学会了撒谎;最后,我们将学习一种读心术,不仅能看懂AI的“集体智慧”,甚至还能预测你的下一步行动。准备好了吗?让我们一起潜入AI的深层意识。 00:00:36 AI进阶之路,当“尖子生”不再需要“课外辅导” 00:05:28 你的AI为什么学不动了?答案可能出乎意料,人多力量大 00:13:04 你的AI助理,如何才能比你更懂你? 00:19:19 AI的“草稿纸”,藏着什么秘密? 00:24:19 AI的“内心戏”,我们终于能看懂了 本期介绍的几篇论文: [CV] Self-Supervised Flow Matching for Scalable Multi-Modal Synthesis [Black Forest Labs] https://arxiv.org/abs/2603.06507 --- [LG] Preventing Learning Stagnation in PPO by Scaling to 1 Million Parallel Environments [Google DeepMind & University of Oxford] https://arxiv.org/abs/2603.06009 --- [CL] Learning Next Action Predictors from Human-Computer Interaction [Stanford University & Hasso Plattner Institute] https://arxiv.org/abs/2603.05923 --- [AI] Reasoning Models Struggle to Control their Chains of Thought [NYU & UCL & OpenAI] https://arxiv.org/abs/2603.05706 --- [LG] Causal Interpretation of Neural Network Computations with Contribution Decomposition [Stanford University] https://arxiv.org/abs/2603.06557
你有没有想过,未来的AI要如何变得更聪明?最新的一些研究告诉我们,答案可能不是一味地堆算力,而是要学会人类的“智慧”。比如,让AI拥有一个能从错误中总结经验的“技能工具箱”;或者像教孩子一样,让它理解规则而不是死记硬背模式;甚至,像一位高明的将军,懂得如何排兵布阵,把好钢用在刀刃上。本期节目,我们就来聊聊这些让AI学会“反思”、“预见”和“布阵”的最新论文,看看真正的智能是如何炼成的。 00:00:38 高手,都是“错”出来的 00:05:41 AI学会举一反三的秘密,换个数字就不认识了? 00:10:57 AI大模型的新兵法,好钢如何用在刀刃上? 00:17:26 让机器人自己“玩”成高手,需要几步? 00:23:29 AI的远见,如何不看细节,反而看得更远? 本期介绍的几篇论文: [AI] EvoSkill: Automated Skill Discovery for Multi-Agent Systems [Sentient & Virginia Tech] https://arxiv.org/abs/2603.02766 --- [LG] Symbol-Equivariant Recurrent Reasoning Models [Johannes Kepler University Linz] https://arxiv.org/abs/2603.02193 --- [LG] DynaMoE: Dynamic Token-Level Expert Activation with Layer-Wise Adaptive Capacity for Mixture-of-Experts Neural Networks https://arxiv.org/abs/2603.01697 --- [RO] Tether: Autonomous Functional Play with Correspondence-Driven Trajectory Warping [University of Pennsylvania] https://arxiv.org/abs/2603.03278 --- [LG] Next Embedding Prediction Makes World Models Stronger [T-Tech] https://arxiv.org/abs/2603.02765
今天我们来聊聊AI世界里那些“反常识”的智慧:为什么“见过世面”的AI不容易遗忘,而“偏科”却成了它发展的隐患?我们不仅会揭秘AI如何通过“预判你的预判”来极致提速,还会探讨为何有时“慢”一点的学习,反而能让AI变得更聪明、更懂变通。最后,我们会发现,解决一个复杂的动画难题,关键可能只是需要为AI发明一种“普通话”。 00:00:31 为什么高手学东西,不容易忘? 00:05:34 AI的加速赛,怎样让聪明的“大脑袋”跑得更快? 00:12:39 AI动画的“普通话”和“方言” 00:17:56 AI智能体:是天才还是“偏科生”? 00:23:02 天下武功,唯快不破?AI训练中的一个“慢”智慧 本期介绍的几篇论文: [LG] Pretrained Vision-Language-Action Models are Surprisingly Resistant to Forgetting in Continual Learning [The University of Texas at Austin & Microsoft Superintelligence] https://arxiv.org/abs/2603.03818 --- [LG] Speculative Speculative Decoding [Stanford University & Princeton University & Together AI] https://arxiv.org/abs/2603.03251 --- [CV] OmniLottie: Generating Vector Animations via Parameterized Lottie Tokens [Fudan University & StepFun & HKU MMLab] https://arxiv.org/abs/2603.02138 --- [AI] How Well Does Agent Development Reflect Real-World Work? [CMU] https://arxiv.org/abs/2603.01203 --- [LG] To Use or not to Use Muon: How Simplicity Bias in Optimizers Matters [New York University] https://arxiv.org/abs/2603.00742
你有没有想过,AI除了会聊天画画,还能做什么更酷的事?本期节目,我们将一口气看到AI能力的多个惊人侧面。从像人一样“脑补”物理世界,到用“笨方法”实现更高效的学习,再到成为物理学家的“科研搭子”,解决真正的科学难题。这些最新论文将刷新你对AI潜力的认知! 00:00:28 AI学会了“脑补”,世界就大不一样了 00:06:08 大模型里的“关系户”,它凭什么吸引了所有注意力? 00:13:10 AI省钱的终极奥义,少就是多 00:18:04 一个“笨方法”,让AI学得更快 00:22:49 AI,从聊天高手到科研搭子 本期介绍的几篇论文: [LG] Latent Particle World Models: Self-supervised Object-centric Stochastic Dynamics Modeling [CMU & UT Austin & Brown University] https://arxiv.org/abs/2603.04553 --- [CL] The Spike, the Sparse and the Sink: Anatomy of Massive Activations and Attention Sinks [New York University] https://arxiv.org/abs/2603.05498 --- [CL] Sparse-BitNet: 1.58-bit LLMs are Naturally Friendly to Semi-Structured Sparsity [Microsoft Research] https://arxiv.org/abs/2603.05168 --- [CL] Replaying pre-training data improves fine-tuning [Stanford University] https://arxiv.org/abs/2603.04964 --- [AI] Solving an Open Problem in Theoretical Physics using AI-Assisted Discovery [Google Research] https://arxiv.org/abs/2603.04735
你是否好奇,为何AI有时会“指鹿为马”?为何它面对难题,内部的神经元反而开始“集体偷懒”?本期节目,我们将通过几篇最新论文,一起给AI的大脑做一次“CT扫描”和“基因测序”,揭示它在感知、学习、思考和效率背后,那些出人意料的底层法则。 00:00:26 人工智能的“阿喀琉斯之踵”,一个关于维度的诅咒 00:05:34 AI绘画进化论,为什么高手不需要“题海战术”? 00:10:02 AI一思考,我们就发笑?不,是神经元在“偷懒” 00:15:44 如何用50倍的效率,给AI做一次“CT扫描”? 00:21:34 AI模型的“不可能三角”,算力、速度与智能 本期介绍的几篇论文: [LG] Solving adversarial examples requires solving exponential misalignment [Stanford University & Aisle] https://arxiv.org/abs/2603.03507 --- [LG] Generalization Properties of Score-matching Diffusion Models for Intrinsically Low-dimensional Data [University of Michigan & Google DeepMind & UC Berkeley] https://arxiv.org/abs/2603.03700 --- [CL] Farther the Shift,Sparser the Representation: Analyzing OOD Mechanisms in LLMs [Rutgers University & Northwestern University & UKP Lab, TU Darmstadt] https://arxiv.org/abs/2603.03415 --- [CL] Compressed Sensing for Capability Localization in Large Language Models [CMU] https://arxiv.org/abs/2603.03335 --- [LG] Why Are Linear RNNs More Parallelizable? [Allen Institute for AI & Rheinland-Pfalzische Technische Universitat] https://arxiv.org/abs/2603.03612
今天,我们要探讨如何让AI从一个只会“动嘴”的聊天伙伴,进化成一个真正“会看、会想、会动手”的智能体。我们会看到,最新论文如何让AI‘开眼看世界’,在脑中建立起预测未来的‘导航系统’,并从海量普通文本中自我启蒙,学会判断好坏。更重要的是,当AI要替我们行动时,它又是如何学会‘三思而后行’,在‘有用’和‘安全’之间找到那条微妙的平衡线呢?准备好了吗?让我们一起探寻AI从‘愣头青’到‘老司机’的进化之路。 00:00:40 AI为什么要“开眼看世界”? 00:07:16 为什么高手都自带“导航系统”? 00:13:19 AI的“行动许可”,它在动手前,先想了什么? 00:19:12 把白开水变成高汤,AI如何从普通文本中学会“好坏” 00:24:47 如何把一个“愣头青”AI,调教成“老司机”? 本期介绍的几篇论文: [CV] Beyond Language Modeling: An Exploration of Multimodal Pretraining [FAIR, Meta] https://arxiv.org/abs/2603.03276 --- [LG] What Capable Agents Must Know: Selection Theorems for Robust Decision-Making under Uncertainty [CMU] https://arxiv.org/abs/2603.02491 --- [LG] Learning When to Act or Refuse: Guarding Agentic Reasoning Models for Safe Multi-Step Tool Use [Microsoft Research] https://arxiv.org/abs/2603.03205 --- [LG] Scaling Reward Modeling without Human Supervision [Harvard University & Cornell University] https://arxiv.org/abs/2603.02225 --- [LG] Safety Training Persists Through Helpfulness Optimization in LLM Agents [UC Berkeley] https://arxiv.org/abs/2603.02229
今天我们不聊模型参数有多大,而是聊如何让AI变得更“会思考”,这种思考方式,有时甚至有些反常识。比如,为什么给AI疯狂“补课”,它反而可能越学越笨?我们还会探讨,如何像一位高明的老师一样引导AI攻克难题,而不是直接灌输答案。更进一步,我们会揭示如何训练AI像个侦探一样,学会“讲道理”地分析代码,以及如何让整个系统学会动态协作,找到最高效的“偷懒”方式。 00:00:35 AI大模型时代,如何花小钱办大事? 00:05:47 给AI“补课”的陷阱,为什么学得越多,它反而越笨? 00:11:37 高手辅导功课,为什么不直接给答案? 00:16:48 让AI学会“讲道理”,代码世界的侦探是怎样炼成的? 00:22:00 让AI学会“省时间”,一种更聪明的快 本期介绍的几篇论文: [LG] Rich Insights from Cheap Signals: Efficient Evaluations via Tensor Factorization [Google DeepMind & University of Michigan] https://arxiv.org/abs/2603.02029 --- [LG] Theoretical Perspectives on Data Quality and Synergistic Effects in Pre- and Post-Training Reasoning Models [University of Southern California & University of California Los Angeles & Google Research] https://arxiv.org/abs/2603.01293 --- [LG] Learn Hard Problems During RL with Reference Guided Fine-tuning [ByteDance Seed & UC Berkeley & CMU] https://arxiv.org/abs/2603.01223 --- [LG] Agentic Code Reasoning [Meta] https://arxiv.org/abs/2603.01896 --- [CL] Learning to Draft: Adaptive Speculative Decoding with Reinforcement Learning [Microsoft Research Asia & Peking University] https://arxiv.org/abs/2603.01639
你有没有想过,一个更聪明的AI,或许需要学会的不是记住一切,反而是“选择性失忆”?本期我们要聊的几篇最新论文,就充满了这样颠覆常识的洞见。我们将一起探索,AI如何从“管住嘴”进化到深入思想的“排毒手术”,如何像顶尖高手一样动态进化自己解决问题的方法论,甚至,如何拥有人类最宝贵的品质之一——知道自己“不知道”的自知之明。 00:00:31 AI“排毒”,是动手术,还是只吃止痛药? 00:04:49 AI的记忆难题,除了死记硬背,还有什么好办法? 00:10:33 你的方法,也需要进化 00:16:14 AI的记忆,竟然是它的负担? 00:21:15 聪明反被聪明误,AI也需要“自知之明” 本期介绍的几篇论文: [LG] Detoxifying LLMs via Representation Erasure-Based Preference Optimization [McGill University & Google DeepMind] https://arxiv.org/abs/2602.23391 --- [LG] Memory Caching: RNNs with Growing Memory [Google Research] https://arxiv.org/abs/2602.24281 --- [LG] EvoX: Meta-Evolution for Automated Discovery [UC Berkeley] https://arxiv.org/abs/2602.23413 --- [CL] Do LLMs Benefit From Their Own Words? [MIT & IBM Research] https://arxiv.org/abs/2602.24287 --- [LG] RewardUQ: A Unified Framework for Uncertainty-Aware Reward Models [ETH Zurich] https://arxiv.org/abs/2602.24040
你有没有想过,为什么AI题刷得越多,反而越容易在简单问题上翻车?这一期,我们将一起潜入AI的内心世界,看看它们是如何陷入“应试教育”的陷阱,又是如何被“剪刀石头布”这样的逻辑死循环给困住的。但更重要的是,我们会发现,科学家们如何通过“读心术”和“记仇本”这样的奇思妙想,教会AI从失败中学习,并找到那条跳出困境的智慧之路。准备好,一场关于AI学习与评估的深度思考,现在开始。 00:00:35 为什么AI刷题越多,第一次答对率反而越低? 00:05:35 AI的“好记性”与“烂笔头” 00:10:06 AI程序员的“应试教育”陷阱 00:14:17 AI世界的“剪刀石头布”难题 00:19:08 机器人教练的“读心术” 本期介绍的几篇论文: [LG] Why Pass﹫k Optimization Can Degrade Pass﹫1: Prompt Interference in LLM Post-training [Singapore University of Technology and Design & University of Maryland] https://arxiv.org/abs/2602.21189 --- [LG] Exploratory Memory-Augmented LLM Agent via Hybrid On- and Off-Policy Optimization [Microsoft Research] https://arxiv.org/abs/2602.23008 --- [LG] ISO-Bench: Can Coding Agents Optimize Real-World Inference Workloads? [Lossfunk] https://arxiv.org/abs/2602.19594 --- [LG] Back to Blackwell: Closing the Loop on Intransitivity in Multi-Objective Preference Fine-Tuning [CMU] https://arxiv.org/abs/2602.19041 --- [RO] TOPReward: Token Probabilities as Hidden Zero-Shot Rewards for Robotics [University of Washington & Amazon] https://arxiv.org/abs/2602.19313
我们给了AI强大的能力,却发现它有时像个“混乱特工”,能为了保守一个秘密烧掉整栋房子。我们以为要给顶尖AI充分的自由,但最新论文却说,给它一份详尽的“任务清单”反而能让它替你赚更多钱。当AI的产出快到我们来不及验证,它的价值又该如何衡量?本期,我们将从几篇最新论文出发,探讨如何驾驭这些能力与心智尚不匹配的强大工具,甚至尝试将AI的“直觉”翻译成我们能懂的“公式”。 00:00:35 请个AI当助理,你放心吗? 00:05:57 AI狂飙,但你的价值正在“空转”? 00:13:56 AI的“直觉”,如何翻译成人类的“公式”? 00:19:45 想让AI替你赚钱?别让它“想太多” 00:24:33 给三维重建,装上一个新引擎 本期介绍的几篇论文: [AI] Agents of Chaos [Northeastern University] https://arxiv.org/abs/2602.20021 --- [AI] Some Simple Economics of AGI [MIT & WashU & UCLA] https://arxiv.org/abs/2602.20946 --- [LG] SymTorch: A Framework for Symbolic Distillation of Deep Neural Networks [University of Cambridge] https://arxiv.org/abs/2602.21307 --- [AI] Toward Expert Investment Teams:A Multi-Agent LLM System with Fine-Grained Trading Tasks [Japan Digital Design, Inc & University of Oxford] https://arxiv.org/abs/2602.23330 --- [CV] VGG-T³: Offline Feed-Forward 3D Reconstruction at Scale [NVIDIA] https://arxiv.org/abs/2602.23361
我们总以为AI的进步靠的是‘大力出奇迹’,但今天我们要聊点更酷的——AI正在学会用‘巧劲儿’。最新论文告诉我们,与其让模型盲目刷题,不如为它铺设一条高效的“语义管道”,甚至教会它像人一样“反思”自己的思考过程。与此同时,AI也正从一个“通才”变身为能帮你管理私人图书馆、设计芯片、甚至用旧零件组装新工具的“超级专家”。准备好了吗?让我们一起看看,AI是如何从‘更强’进化到‘更聪明’的。 00:00:35 大力出奇迹?人工智能的另一条捷径 00:06:08 给你一座私人图书馆,还配一个秒懂你的图书管理员 00:12:19 造AI,有了一本“宜家说明书”? 00:17:49 AI不止会聊天,它还能设计芯片了? 00:25:09 教AI反思,比喂它知识更重要 本期介绍的几篇论文: [LG] Semantic Tube Prediction: Beating LLM Data Efficiency with JEPA [Atlassian & NYU & Brown] https://arxiv.org/abs/2602.22617 --- [IR] DS SERVE: A Framework for Efficient and Scalable Neural Retrieval [UC Berkeley & University of Illinois Urbana–Champaign] https://arxiv.org/abs/2602.22224 --- [CL] dLLM: Simple Diffusion Language Modeling [UC Berkeley & UIUC] https://arxiv.org/abs/2602.22661 --- [LG] ArchAgent: Agentic AI-driven Computer Architecture Discovery [Google & UC Berkeley] https://arxiv.org/abs/2602.22425 --- [LG] Mirroring the Mind: Distilling Human-Like Metacognitive Strategies into Large Language Models [Seoul National University] https://arxiv.org/abs/2602.22508
今天我们要聊聊,如何让AI变得更快、更聪明,甚至更“会过日子”和更“懂合作”。我们将一起探索,怎样用一种巧妙的修剪方法,让AI的信息地图建造速度提升十几倍;又如何给AI大脑动个“小手术”,打通它的多步推理思路。我们还会发现,AI不仅能看着YouTube的“野生”视频自己学会开车,还能像个精明的项目经理一样,把预算花在刀刃上。最后,我们将见证AI如何在一个游戏世界里,第一次学会“换位思考”,理解一个由我们共同构成的现实。 00:00:40 你的AI为什么“反应慢”?问题可能出在建图上 00:06:29 给AI大脑动个“小手术” 00:11:59 AI学车的新思路,让YouTube当免费教练 00:17:32 AI的省钱之道,把钱花在刀刃上 00:22:37 AI学会了“换位思考”,世界会有什么不同? 本期介绍的几篇论文: [IR] PiPNN: Ultra-Scalable Graph-Based Nearest Neighbor Indexing [UMD & Google Research] https://arxiv.org/abs/2602.21247 --- [LG] Interleaved Head Attention [Meta & UT Austin & MIT] https://arxiv.org/abs/2602.21371 --- [CV] Learning to Drive is a Free Gift: Large-Scale Label-Free Autonomy Pretraining from Unposed In-The-Wild Videos [Applied Intuition & Stanford University & UC Berkeley] https://arxiv.org/abs/2602.22091 --- [CL] Budget-Aware Agentic Routing via Boundary-Guided Training [University of Cambridge & M365 Research, Microsoft] https://arxiv.org/abs/2602.21227 --- [CV] Solaris: Building a Multiplayer Video World Model in Minecraft [New York University] https://arxiv.org/abs/2602.22208
与播客爱好者一起交流
添加微信好友,获取更多播客资讯
播放列表还是空的
去找些喜欢的节目添加进来吧