筛选结果
48 篇排序
1:25:01EN → 简体中文Gary Marcus:深度学习模仿智能,却并不理解嘉宾 Gary Marcus · 主持 Lex Fridman在 Lex Fridman 的播客上,这位纽约大学荣休教授、《重启 AI》(Rebooting AI)的合著者认为,感知与推理需要不同的工具;在机器能够表示伤害、瓶子和意图之前,可信的 AI 不可能实现。
1:17:29EN → 简体中文算法已经更好,诊所却还在用 1967 年的标准嘉宾 Regina Barzilay · 主持 Lex FridmanMIT 的 Regina Barzilay 讲,深度学习(deep learning)为什么已经胜过乳腺癌的“密度”评分,而数据、法律责任和激励机制为什么仍然推不动改变。
1:59:49EN → 简体中文智能不是楼的高度:François Chollet 谈 Keras、深度学习,以及为什么通用人工智能(AGI)不会爆发式崛起嘉宾 François Chollet · 主持 Lex FridmanKeras 作者在 Lex Fridman 的节目里说,智能是大脑、身体与问题相互作用的产物;科学早已是一个会递归自我改进的超人系统,产出却只呈线性增长;而近期真正的危险是信息流。
1:44:11EN → 简体中文一块 GPU 就够了:Jeremy Howard 谈如何教深度学习嘉宾 Jeremy Howard · 主持 Lex Fridmanfast.ai 创始人在与 Lex Fridman 的对话中说,真正决定深度学习(deep learning)能否帮到人的,是领域专家,再多的论文或 TPU 都无济于事。
1:10:58EN → 简体中文让谷歌开源 TensorFlow 的 Hadoop 教训嘉宾 Rajat Monga · 主持 Lex Fridman2015 年 11 月,谷歌把自己内部的核心机器学习系统放上了 GitHub。Rajat Monga 解释,像谷歌这样体量的公司为什么要把最难复制的那部分交出去。
1:46:01EN → 简体中文AlphaStar 击败 TLO 那晚,团队去了酒吧嘉宾 Oriol Vinyals · 主持 Lex FridmanOriol Vinyals 谈 DeepMind 如何把《星际争霸 II》变成一个序列问题、搭建智能体联盟,以及为什么他仍然不肯说这款游戏已经被解决。
1:08:37EN → 简体中文蒙特利尔的一家酒吧,然后是一夜写代码:Ian Goodfellow 谈 GAN 为何奏效嘉宾 Ian Goodfellow · 主持 Lex FridmanLex Fridman 与 Ian Goodfellow 对谈,聊 2014 年生成对抗网络(GAN)的发明、作为安全问题的对抗样本(adversarial examples),以及机器要像人一样掌握一个事实,还需要建造什么。
1:20:21EN → 简体中文爱因斯坦曾是五名博士生里最差的一个嘉宾 Tomaso Poggio · 主持 Lex FridmanMIT 神经科学家 Tomaso Poggio 谈智能为何是科学中最深的难题,以及大脑仍有什么要教给机器的东西。
1:19:59EN → 简体中文Schmidhuber 的赌注:真正的元学习改写的是学习者本身,而不只是最后一层嘉宾 Juergen Schmidhuber · 主持 Lex FridmanJürgen Schmidhuber 向 Lex Fridman 讲述,哥德尔机、把压缩当作科学、Power Play、LSTM 和世界模型,如何都源于一个少年时的计划:把一个人的好奇心放大,造出一个比他本人强得多的物理学家。
42:45EN → 简体中文单腿机器人只凭“这个更好”学会了后空翻嘉宾 Pieter Abbeel · 主持 Lex FridmanPieter Abbeel 对 Lex Fridman 说明,为什么偏好比较、自我对弈、十分钟示范,以及杂乱的模拟器组合,是深度强化学习(RL)的切实路径。
54:03EN → 简体中文教科书略去了学习中最难的那部分嘉宾 Vladimir Vapnik · 主持 Lex FridmanVladimir Vapnik 在与 Lex Fridman 的对谈中认为,智能的核心工作是发明能缩小可容许函数集的谓词,在一大堆数据上拟合规则做不到这一点。
42:18EN → 简体中文成年人能改变去年的想法,网络却做不到嘉宾 Yoshua Bengio · 主持 Lex FridmanYoshua Bengio 对 Lex Fridman 说,深度学习缺的那几块,在于信用分配、训练目标,以及从经典 AI 借来的一种组合性,靠堆出一万层网络补不上。
15:01ENDeep Learning Is a System 1 Machine: Daniel Kahneman on What AI Still Lacks嘉宾 Daniel Kahneman · 主持 Lex FridmanThe Nobel-winning psychologist tells Lex Fridman how his fast-and-slow thinking framework maps onto deep learning, and why driving among pedestrians may require understanding people.
54:33ENVision or Lidar, Human or Machine: Lex Fridman's 2019 Map of the Self-Driving Race嘉宾 Self-Driving Cars · 主持 Lex FridmanThe opening lecture of MIT's 6.S094 course surveys 2018's milestones, the industry's optimistic predictions, and the beliefs about driving and humans that will decide who wins.
10:16ENReasoning Without Logic: Yann LeCun on Teaching Neural Networks to Think嘉宾 Yann Lecun · 主持 Lex FridmanYann LeCun explains why he believes neural networks can reason, and what memory, recurrence and continuous representations would need to look like.
17:21ENHow Alexa Learned to Hear From Across the Room: Rohit Prasad on Far-Field Speech and Intent嘉宾 Rohit Prasad · 主持 Lex FridmanAlexa's head scientist recalls the early days of the Echo, when far-field speech recognition seemed unsolvable and there were no customers to collect data from.
1:08:06ENDeep Learning in One Lecture: From Perceptrons to AutoML and the Limits of Learning嘉宾 Deep Learning Basics · 主持 Lex FridmanLex Fridman's opening lecture for MIT's 6.S094 course surveys why deep learning took off, how neural networks learn, and where the technique still falls short.
24:08ENIntelligence Is Not Just Prediction: David Ferrucci on Explanation, Persuasion and Shared Meaning嘉宾 David Ferrucci · 主持 Lex FridmanThe lead of IBM's Watson team argues that predicting well is only half of intelligence, and that being understood by others is the other half.
10:34ENWhy Deep Learning Needs Rules: Chollet on Interpolation, Abstraction and Hybrid AI嘉宾 François Chollet · 主持 Lex FridmanFrançois Chollet explains why neural networks only interpolate between points they have seen, and why the best AI systems pair them with symbolic rules.
52:28ENRay Kurzweil on Hierarchies, the Neocortex and Merging with AI嘉宾 Ray Kurzweil · 主持 Lex FridmanIn an MIT AGI course lecture, Kurzweil connects the history of neural networks to his hierarchical theory of the neocortex, language understanding at Google, longevity, jobs and existential risk.
1:00:15ENCircuit Search, Hindsight and Self-Play: Ilya Sutskever's Case for Where Learning Is Going嘉宾 Ilya Sutskever · 主持 Lex FridmanIn an MIT AGI lecture, OpenAI's co-founder explains why deep learning works, then walks through meta-learning, hindsight experience replay, sim-to-real transfer, self-play and learning goals from human feedback.
11:59ENFrom a Side Project to TensorFlow's Core: François Chollet on How Keras Came to Be嘉宾 François Chollet · 主持 Lex FridmanThe creator of Keras recalls why he built it in 2015 around recurrent networks, how it moved onto TensorFlow, and what TensorFlow 2.0 means for its design.
26:45ENStart Small, Build Habits, Pick Your Team: Andrew Ng's Practical Advice for Learning Deep Learning嘉宾 Andrew Ng · 主持 Lex FridmanIn a clip from the Lex Fridman Podcast, Andrew Ng explains how to begin in deep learning, how to study effectively, and how to choose a career path.
51:18ENEngineering Intelligence: Lex Fridman's Opening Lecture for MIT's AGI Course嘉宾 MIT AGI · 主持 Lex FridmanLex Fridman lays out an engineering-first approach to artificial general intelligence, previews the course's projects and guest speakers, and frames deep learning's limits.







