
[人人能懂AI前沿] 从协同进化、无意泄露到经济学路由给机器人请个“高人”当教练,它就能更快出师吗?你越是强调一个秘密,AI助理反而越容易通过“微表情”泄密?面对眼花缭乱的AI模型,怎样才能做出最“划算”的选择?我们又该如何把你电脑里那些只可意会的隐形操作,变成一本AI也能看懂的“武功秘籍”?本期节目,我们将透过几篇最新论文,一起探索如何让AI学会更高效地行动、更安全地协作,以及更聪明地为我们当好管家。 00:00:33 给机器人请个“高人”当教练 00:06:16 你的AI助理,可能是个藏不住事的“大嘴巴”? 00:11:00 你的下一个AI,需要一个“划算”计算器 00:17:13 如何让AI“学徒”早出师? 00:21:37 你的电脑,藏着一本“隐形说明书” 本期介绍的几篇论文: [AI] EXIMO: VLM Guided Exploration of VLA Policies [Google DeepMind] https://arxiv.org/abs/2608.19891 --- [LG] Inadvertent Context Leakage in Language Models [Meta Superintelligence Labs & UC Berkeley] https://arxiv.org/abs/2608.19857 --- [AI] Pandora's AI Model Routing Box: Efficient Allocation with Costly Value Estimation [Google DeepMind] https://arxiv.org/abs/2608.20316 --- [AI] MidTool: Mid-training Data Synthesis for Agentic Tool Use [University of Washington & Snowflake] https://arxiv.org/abs/2608.20314 --- [CL] Inducing Task Models from Computer-Use Traces [Stanford University & CMU] https://arxiv.org/abs/2608.20319
[人人能懂AI前沿] 解码大脑、重塑逻辑与应对“祸不单行”你是否想过,AI要如何才能像武林高手一样“左右互搏”,自己给自己出题,实现无限成长?当它学习一项新技能时,又要如何避免像我们一样,一紧张就把基本功忘得一干二净?更神奇的是,我们还将看到AI如何不靠开颅手术,就能精准“读懂”我们大脑里的句子。本期节目,我们将通过几篇最新论文,一起探寻AI世界里关于学习、成长与解决复杂问题的非凡智慧。 00:00:33 当AI学会“读心”,我们离未来还有多远? 00:07:00 突破成长天花板,AI如何学会“左右互搏”,做自己最好的老师? 00:12:39 告别“狗熊掰棒子”式的努力,从一台AI机器手的进化,看高手的“底层能力”构建 00:16:14 破局“好与快”的死结,从猜答案到重塑底层逻辑的认知飞跃 00:21:06 当麻烦“祸不单行”时,我们该如何破局?,,来自前沿AI算法的生存智慧 本期介绍的几篇论文: [CL] Accurate Decoding of Natural Sentences from Non-Invasive Brain Recordings [Meta AI & Université PSI] https://arxiv.org/abs/2608.18114 --- [CL] SPADE: Self-Play in Adaptive Synthetic Executable Environments [University of Washington & Northeastern University & CMU] https://arxiv.org/abs/2608.19197 --- [RO] ADEPT: Accelerating Dexterity via Pre-Training and Post-Training using Reinforcement Learning [NVIDIA] https://arxiv.org/abs/2608.19182 --- [AI] Coupled-cluster molecular properties across the main group that extrapolate beyond training size [MIT] https://arxiv.org/abs/2608.18346 --- [LG] Continuous-Time Reinforcement Learning for Controlled Hawkes Jump-Diffusions [UC Berkeley] https://arxiv.org/abs/2608.19151
[人人能懂AI前沿] 从反刍记忆、辩证学习到分治与预见今天,我们不聊堆算力的“大力出奇迹”,而是要探索几条让AI变得更智慧、更可靠的巧妙路径。我们会看到,一个简单的“反刍”机制,如何让AI不再健忘;一场内部“辩论赛”,又如何教会它诚实;“专家分工”的智慧,怎样让它在变强的同时还更省钱。最后,我们还会探究AI是如何学会像高手一样“抬头看路”地做决策,甚至在数学领域领悟“功夫在题外”的道理。准备好了吗?让我们一起揭开这些最新论文背后的绝妙构思。 00:00:37 AI的“反刍”,一个让它更聪明的简单魔法 00:05:10 如何让AI变得更聪明,同时还不变坏? 00:09:50 AI 进化新思路,从“大力出奇迹”到“聪明分工” 00:14:55 高手决策的秘密,既要埋头拉车,又要抬头看路 00:20:36 AI做数学,功夫在诗外 本期介绍的几篇论文: [LG] Recirculation [Google DeepMind] https://arxiv.org/abs/2608.17981 --- [LG] Debate Training Reduces Reward Hacking in RLAIF [Google DeepMind] https://arxiv.org/abs/2608.17776 --- [CV] MoE-ViE: Mixture of Experts Vision Encoder for Efficient Image and Video Understanding [Meta] https://arxiv.org/abs/2608.17402 --- [LG] Q-Learning With World Models [Stanford University & Peking University] https://arxiv.org/abs/2608.17163 --- [AI] The Problem Is the Problem: Towards Scalable Mathematical Discovery [CMU] https://arxiv.org/abs/2608.16977
[人人能懂AI前沿] 从跨界工具、群体动力学到长时记忆与元认知你有没有想过,聪明的AI也会犯傻,甚至像个没头脑的实习生?本期节目,我们就来聊聊如何让AI变得更“靠谱”。我们将一起看看,科学家们如何用AI工具去解决古老的数学难题,如何洞悉AI群体的“集体意识”,是会变得更聪明还是更固执,以及如何教会AI拥有一个好记性,并像人一样学会“反思”自己。 00:00:28 给你一把新扳手,拧紧一颗老螺丝 00:05:56 AI的“集体意识”,乌合之众还是三个臭皮匠? 00:10:51 如何才能拥有一个好记性? 00:15:37 为什么聪明的AI,干起活来却像个“没头脑”? 00:21:07 给AI立规矩,为什么不能靠“死命令”? 本期介绍的几篇论文: [LG] Improving the matrix multiplication exponent with modern optimization and AlphaEvolve [Google DeepMind] https://arxiv.org/abs/2608.16884 --- [AI] Physics of Agents: Statistical Mechanics Predicts Collective Behavior of AI Agents [Stanford University & UC Santa Barbara] https://arxiv.org/abs/2608.16578 --- [LG] Proteus: Incremental Memory Activation for Long-Context Sequence Modeling [Mila & Google] https://arxiv.org/abs/2608.16844 --- [CL] How Do Agents Fail on AutoResearch: End-to-End Diagnostic Evaluation on 100 Real-World Frontier Research Tasks [Prentis AI] https://arxiv.org/abs/2608.14905 --- [CL] CAPO: Constraint-Aware Prompt Optimization for LLM Agents [Microsoft] https://arxiv.org/abs/2608.16068
[人人能懂AI前沿] 从统一路径、模块涌现到元认知鸿沟今天我们来当一回AI世界的侦探,看看AI的“黑箱”里都藏着哪些秘密。我们将揭开AI绘画两大流派的统一秘诀,看看AI的大脑里是不是也分出了“文科”和“理科”部门。接着,我们会分辨AI是在“真思考”还是在“表演思考”,并学习它如何为未知游戏自建一个“数字孪生”。最后,再看看科学家如何给这个聪明的“大脑”进行一次外科手术级的精准“瘦身”,让它跑得更快更好。 00:00:31 AI绘画高手,为何在“半路”上吵翻了天? 00:06:03 AI的大脑里,也分“文科”和“理科”吗? 00:10:21 你是在真思考,还是在表演思考? 00:15:38 如何像高手一样,玩一把没说明书的游戏? 00:20:16 AI绘画的“火候”,高手与庸才的分野 本期介绍的几篇论文: [LG] Designing Reinforcement Learning for Diffusion Models: A Unified Path-Space View [Peking University & ByteDance Seed] https://arxiv.org/abs/2608.14430 --- [AI] Modular Cognitive Architecture Emerges in Large Language Models [MIT] https://arxiv.org/abs/2608.13567 --- [CL] Amplified Does Not Mean Predictive: Reasoning Behaviors in Thinking Models [CMU] https://arxiv.org/abs/2608.13760 --- [AI] Twin: Playing an Unknown Game with a Test-Time Digital Twin [Yeshiva University & Stanford University & Cornell University] https://arxiv.org/abs/2608.14490 --- [LG] Adversarial Learning of Classifier-Free Guidance Schedules [Google & Google DeepMind] https://arxiv.org/abs/2608.14038
[人人能懂AI前沿] AI的品味、情绪与边界感我们总觉得AI变得更强,就是模型更大、算力更猛,但今天我们要聊点不一样的。最新几篇论文告诉我们,真正的智能升级,是教会AI拥有科学家的“品味”,甚至赋予它类似人类的“情绪”来感知对错。同时,我们还要用一点小小的“随机”来防止它学会“耍滑头”,并在一场终极“摸底考”中,看清它距离人类顶尖黑客到底还有多远。准备好了吗?让我们一起看看,AI如何被塑造出更深邃的智慧。 00:00:33 AI 会“品”,科学大不同 00:07:02 你的AI有“情绪”了,而且这决定了它的智商 00:13:00 如何防止你的AI员工「耍滑头」? 00:18:15 人工智能摸底考,为什么黑客的饭碗暂时还很稳? 00:24:19 AI法官的“内心戏”,一个比准确率更重要的指标 本期介绍的几篇论文: [LG] Training AI Scientists to Replicate Research [Inherent] https://arxiv.org/abs/2608.13331 --- [AI] Emotion2Skill: Model-Internal Emotion Signals for Adaptive Skill Selection and Evolution [University of Science and Technology of China & University of Oxford & University of Arizona] https://arxiv.org/abs/2608.09248 --- [LG] Rubric Dropout: A Simple Way to Mitigate Reward Hacking in Rubric-as-Reward RL [Scale AI & University of Arizona] https://arxiv.org/abs/2608.11669 --- [AI] The Next Challenge for Agentic Cybersecurity: A Realistic, Contamination-Free Reverse Engineering Benchmark [Columbia University & UC Berkeley] https://arxiv.org/abs/2608.11469 --- [AI] Jagged Judges: Epistemic Stability Under Silence, Pressure, and Persistence [Meta Superintelligence Labs] https://arxiv.org/abs/2608.12645
[人人能懂AI前沿] 沉默的思考者、诚实的学徒与家族里的“内鬼”你有没有想过,AI在给你答案之前,它的大脑里到底发生了什么?本期我们来聊聊AI几种奇特的“思考术”:有的AI学会了更省钱的“默算”,有的则像一个项目经理,懂得把复杂任务拆解成一个个小技能包。同时,我们也会揭示一个惊人漏洞——AI家族里的“小弟”是如何出卖“大哥”的商业机密;以及,AI学徒又该如何在一个绝对安全的环境里,把自己训练成“股神”。这些最新论文,正在重新定义AI的智慧、效率与安全边界。 00:00:35 AI的“默算”能力,更聪明,还是更经济? 00:05:47 AI写论文?不,它在学习一种更重要的能力 00:11:09 你家AI的“悄悄话”,正在被隔壁“笨小孩”出卖 00:17:00 AI当学徒,能把自己教会成股神吗? 00:22:44 你的AI闯了祸,到底该谁来背锅? 本期介绍的几篇论文: [AI] BDH-CQ: In-Context Learning with Recurrent Latent Reasoning [B Engdahl, A Kosowski, J Chorowski, Z Stamirowska…] https://arxiv.org/abs/2608.09888 --- [CL] Spark-to-Paper: End-to-End Research Paper Generation as a Composable Skill [Vast Intelligence Lab & University of Technology Sydney] https://arxiv.org/abs/2608.11924 --- [AI] Stealing Reasoning Traces from Proprietary LLM APIs [MATS Research & ELLIS Institute Tübingen & AI Security Company] https://arxiv.org/abs/2608.09867 --- [CL] AQuA: Recursively Self-Improving Quantitative Trading Research Agents [Princeton University & Ant Group] https://arxiv.org/abs/2608.12841 --- [AI] Legal Responsibilities Using Autonomous Agents For Artificial Intelligence [ChiTek-i AS] https://arxiv.org/abs/2608.08022
[人人能懂AI前沿] 拼图高手、师徒搭档与密室玩家AI画画写代码,怎样才能告别蛮力,像高手一样把力气用在刀刃上,又像学徒一样得到名师指点,快速开窍呢?它的学习过程到底是充满“顿悟”的跳跃,还是一分耕耘一分收获的苦功?更进一步,当规则完全未知时,AI能像我们玩密室逃脱一样,自己摸索出世界的法则吗?本期节目,我们就从四篇最新论文出发,一起探寻AI从“聪明”走向“智慧”的秘密。 00:00:31 生成AI的“节拍器”,如何把算力用在刀刃上? 00:06:13 AI当码农,如何从“笨徒弟”进化成“老师傅”? 00:12:35 AI学习的秘密,顿悟与苦功,本来就是一回事 00:19:14 AI的下一个考场,在规则未知的世界里摸索 00:24:27 AI养娃,要从胎教开始 本期介绍的几篇论文: [LG] The data geometry of masking diffusion: Certified-optimal schedules via unmasking growth complexity [M J. Wainwright, MIT] https://arxiv.org/abs/2608.13520 --- [LG] CAKE: Compiler-Agent Co-Design for Frontier Kernel Evolution [Z Ye, Y Huang, H Jin, B Hou… (NVIDIA & CMU)] https://arxiv.org/abs/2608.12629 --- [LG] Neural Quadratic Forms: A Unified Minimal Model for Sudden Learning and Scaling Laws [L Ziyin, Y Xu, T Poggio, I Chuang (MIT & EPFL)] https://arxiv.org/abs/2608.13335 --- [LG] DiG-bench: Discovery in Games [R M. Battleday, K Sandbrink, J Cullen-Drohan, Z Yan… (Thinking About Thinking)] https://arxiv.org/abs/2608.12593 --- [LG] Synthetic Persona Pretraining: Alignment from Token Zero [J Minder, V Moskvoretskii, R Singhal, D Jiao,… (EPFL)] https://arxiv.org/abs/2608.13482
[人人能懂AI前沿] 从全栈优化、信息悖论到模拟器坍塌你有没有想过,让人工智能变聪明的秘诀,可能不是“更多”,而是“更巧”?本期我们要聊的几篇最新论文,就充满了这种“反常识”的智慧:从把效率从细节里“省”出来,到警惕信息太丰富反而让AI“变笨”的悖论。我们还会看到,一个“完美”的陪练为何会带出最差的学生,以及如何通过精心呵护AI的“童年”,来预测它未来的潜力。准备好,让我们一起在这些看似矛盾的发现中,窥见AI的未来。 00:00:32 省出来的效率,才是真本事 00:07:29 AI的“富贵病”,为什么信息越多,它反而越“笨”? 00:11:44 你的AI陪练,正在让你变傻 00:17:46 人工智能的“童年”里,藏着未来的密码 00:22:55 AI瘦身术,从“一刀切”到“看人下菜”的智慧 本期介绍的几篇论文: [LG] Dion3: Full-Stack Orthogonal Updates [New York University & Princeton University & NVIDIA] https://arxiv.org/abs/2608.11612 --- [CL] Information Abundance Paradox: Long-Context Training Undermines Parametric Knowledge [Johns Hopkins University] https://arxiv.org/abs/2608.12218 --- [CL] One Frozen Simulator Is Not Enough: Simulator Collapse in Multi-Agent RL [Northeastern University & New York University & UC Berkeley] https://arxiv.org/abs/2608.12253 --- [LG] Small-Scale Experiments: Are We There Yet? [FAIR at MSL Meta & New York University] https://arxiv.org/abs/2608.11859 --- [LG] SoftWater: Class-Aware Rate Allocation for Softmax Quantization [MIT] https://arxiv.org/abs/2608.12026
[人人能懂AI前沿] 效率的代价、思维的几何与价值观的简化你有没有想过,我们每天都在用的AI,在那些看不见的地方,正在发生什么?本期我们将通过几篇最新论文,一起去看看AI华丽大厦地基下的“裂缝”,潜入它用于思考的“秘密厨房”。我们还会探讨如何为它装上一个检测内心矛盾的“逻辑测谎仪”,并警惕我们是怎样在不经意间,把复杂的“人类价值观”简化成了一道危险的选择题。 00:00:30 AI大模型,那些藏在基座里的“裂缝” 00:05:30 你的AI在说谎吗?我们迎来了一个“逻辑测谎仪” 00:11:39 AI的“心口不一”,它在哪以及为什么在那思考? 00:16:43 AI的价值观,正在被简化成一道选择题 00:21:45 让机器人拥有“故事感”的记忆 本期介绍的几篇论文: [CL] Cracks in the Foundation: Seemingly Minor Architectural Choices Impact Long Context Extension [Ai2 & CMU] https://arxiv.org/abs/2608.10296 --- [AI] How to Verify Consistency of Probabilistic Claims [EPFL & Université de Montréal] https://arxiv.org/abs/2608.11181 --- [CL] Off-Axis, On Purpose: Where a Transformer Computes Concepts and Why it Does So [University of Washington] https://arxiv.org/abs/2608.10251 --- [AI] Toward a Theory of Value in AI Alignment [Google Research & UCLA & Google DeepMind] https://arxiv.org/abs/2608.10327 --- [CV] GESTO: Human-Centric Spatio-Temporal Memory for Reasoning in Dynamic Scenes [KTH Royal Institute of Technology & University of Stuttgart] https://arxiv.org/abs/2608.10886
[人人能懂AI前沿] AI如何学会了不浪费、不盲动、不瞎忙?今天,我们来聊聊如何让AI不再只靠“大力出奇迹”,而是学会更聪明地工作。我们会看到,AI如何学会“继承”自己的思考,不再用后即焚;又如何像个聪明的导演,把算力“增援”到最关键的地方。我们还会发现,机器人如何掌握了快慢有度的“节奏感”,以及一个好的系统为何要懂得“聪明的懒惰”。这几篇最新论文,将带我们一窥AI从“野蛮生长”到“精耕细作”的进化之路。 00:00:32 让AI告别“用后即焚”的思考模式 00:05:29 AI解题新思路,如何把一份算力,掰成八瓣花? 00:10:45 机器人也懂的“快慢之道” 00:15:42 成大事者,为什么都懂得“懒惰”的艺术? 00:21:29 给AI一盒乐高,让它自己搭出新世界 本期介绍的几篇论文: [AI] Full-bandwidth transformer [Johns Hopkins University & Princeton University & Microsoft] https://arxiv.org/abs/2608.08888 --- [AI] Thought-Level Beam Search for Reasoning [Princeton University & MIT & Meta AI] https://arxiv.org/abs/2608.08020 --- [RO] SpeedTuning: Speeding Up Policy Execution with Lightweight Reinforcement Learning [Stanford University] https://arxiv.org/abs/2608.09138 --- [LG] Beyond Binary: Continuous State Optimization with Graph-Structured Objectives [Google Research & Tel Aviv University] https://arxiv.org/abs/2608.09366 --- [LG] Idea Search: Guiding Tree Search with Ideas to Explore Diverse Scientific Methods [California Institute of Technology & Google Research] https://arxiv.org/abs/2608.08958
[人人能懂AI前沿] 从模拟实践、耦合定律到裁判分片今天我们来聊聊如何把聪明的AI,变成一个真正可靠的专家。我们会看到,AI要像医生一样去“实习”才能成长,而训练它需要一张全新的“地图”。我们还将揭开手机AI突然“变笨”的秘密,并告诉你一个简单方法,让AI裁判不再“偷懒”。这几篇最新论文,将刷新你对AI如何学习和工作的认知。 00:00:27 AI医生实习记,高手是怎么炼成的? 00:05:15 大模型训练,高手手里的那张新地图 00:11:00 你的手机AI,为什么会突然变笨? 00:17:12 AI裁判也会“偷懒”?一个简单的办法让它更靠谱 00:22:18 大模型瘦身指南,你以为的“闲职”,其实是“关键先生” 本期介绍的几篇论文: [AI] ResidencyRL: Reinforcement Learning in Simulated Clinical Environments [Google DeepMind] https://arxiv.org/abs/2608.07418 --- [CL] Skaling: Chinchilla's Exponents Meet Kaplan's Coupling [FAIR at Meta] https://arxiv.org/abs/2608.07222 --- [LG] Quantization Damage Is Multiplicative, Not Additive [Holistic AI] https://arxiv.org/abs/2608.06564 --- [LG] Sharding Prevents LLM Oversight Failures and Adversarial Exploitation [CMU] https://arxiv.org/abs/2608.06422 --- [LG] The Sparsity Whisperer [MIT] https://arxiv.org/abs/2608.06630
[人人能懂AI前沿] 揭秘AI的执行力、工作流与反思力AI是如何学会“成事”的?本期节目,我们将看到,AI如何通过处理办公室杂活,竟然领悟了解决复杂问题的底层心法。我们还会揭秘一套神奇的“管家系统”,看它如何防止聪明的AI在长任务中掉链子。但与AI聊得太久,为何反而会陷入危险的“妄想旋涡”?最后,当任务完成,AI又是如何精准地判断出,哪一步才是真正的功臣? 00:00:29 成事的底层心法,AI学会了,我们呢? 00:06:12 你的AI为什么总掉链子?因为它缺个好管家 00:12:07 为什么和AI聊得越久,就越危险? 00:18:45 功劳怎么算?AI学会了“动态归因” 00:25:15 AI生成,从“万里长征”到“瞬间移动” 本期介绍的几篇论文: [AI] Post-Training on Office Work Improves Software Engineering: A Behavioral Account of Cross-Domain Transfer [Surge AI] https://arxiv.org/abs/2608.01604 --- [CV] LongHorizon-Harness: Advancing Long-Horizon Agents for Real-World Tasks [DreamX Team, Alibaba Group] https://arxiv.org/abs/2608.01964 --- [CL] DelusionEval: Measuring Delusion-Linked Behaviors in AI Chatbots [Stanford University] https://arxiv.org/abs/2608.05004 --- [AI] AgentOPSD: Recursive Self-Distillation for Agentic Reinforcement Learning [Tsinghua University & Zhejiang University] https://arxiv.org/abs/2608.05987 --- [LG] Beckmann Transport Models: From Autonomous Flows to One-Step Maps [Harvard University & Capital Fund Management & University of Oxford] https://arxiv.org/abs/2608.01692
[人人能懂AI前沿] 从数字彩排、戴镣起舞到跳出像素格今天,我们不聊AI有多聪明,而是聊它如何变得更“懂事”、更“实用”。本期节目,我们将透过几篇最新论文,看看AI如何用83亿虚拟人格为产品进行“数字彩排”。同时,我们也会探讨AI如何学会在现实世界的重重限制下“戴着镣铐跳舞”。最后,我们将一窥AI如何将理解、创造和编辑融为一体,跳出二维像素的禁锢,成为真正强大的三维世界“造物主”。 00:00:32 在数字世界里,我们如何“彩排”未来? 00:06:19 你的AI员工,能戴着镣铐跳舞吗? 00:10:58 数字世界的“造物主”工具箱 00:16:15 跳出像素格,才能看见真实的三维世界 00:21:11 机器人偷师记,它怎么学会了我们干的活? 本期介绍的几篇论文: [AI] MatrAIx: Simulating the World with 8.3 Billion Persona Agents [MatrAIx] https://arxiv.org/abs/2608.04205 --- [AI] Permission Denied: Policy-Graded Evaluation of Coding Agents in Hardened Environments [Accomplish AI] https://arxiv.org/abs/2608.02670 --- [CV] Hunyuan3D-Buffalo 1.0: A Unified Multimodal Model for Scalable 3D Generation, Understanding, and Editing [Tencent Hunyuan] https://arxiv.org/abs/2608.02711 --- [CV] InfiniSplat: Implicit Gaussian Decoding for Large-Baseline Monocular View Synthesis [Zhejiang University] https://arxiv.org/abs/2608.02437 --- [RO] Ego2Robot: Scalable Robot Data Synthesis from Egocentric Human Data [Qwen Team & Renmin University of China] https://arxiv.org/abs/2608.02580
[人人能懂AI前沿] 从自我一致、层级远见到极简对齐你是否也好奇,为什么AI时而是个观点摇摆的“墙头草”,时而又像个只顾眼前、缺乏远见的“短视司机”?本期节目,我们将通过四篇最新论文,揭示AI如何学会拥有稳定的观点和深谋远虑的智慧。我们还将发现,解决复杂问题,有时最简单的数据“对齐”就能力压千钧;甚至,善意添加的正确数据,反而会变成“毒害”AI的糖衣炮弹。准备好,让我们一起深入AI的“思想内核”! 00:00:33 如何让AI不再当“墙头草”? 00:05:34 AI进化新思路,从“下一步”到“下一站” 00:10:09 预测未来,与其“魔改”,不如“对齐” 00:16:05 好心办坏事,为什么正确的数据也会“毒害”人工智能? 00:21:55 为什么最优的健康方案,可能不是最可靠的选择? 本期介绍的几篇论文: [CL] Position: It's Time to Optimize LLMs for Self-Consistency [MIT] https://arxiv.org/abs/2608.05188 --- [CL] Hierarchical Latent Prediction for Language Models [Microsoft Research & University of Texas at Austin] https://arxiv.org/abs/2608.05806 --- [LG] Align-RAG: Alignment Is All You Need for TSFM In-Context Learning [Stanford University & Amazon] https://arxiv.org/abs/2608.05571 --- [LG] Optimal Rates for Learning with Monotone Adversaries [Stanford University] https://arxiv.org/abs/2608.06337 --- [LG] Quality Diversity for Reliable Data Driven Time-Use Optimization [Adelaide University] https://arxiv.org/abs/2608.05230