Published Jan 8, 2022

Eric Jang on Robots Learning at Google and Generalization via Language

Eric Jang delves into the groundbreaking impact of language in advancing robotic learning and generalization at Google, emphasizing the superiority of language-based conditioning over video methods. He explores challenges and innovations in robotics, from semantic grasping to meta and reinforcement learning, showcasing the evolution of intelligent robotic manipulation and task proficiency.

Episode Highlights

Topics covered

Episode Highlights

Related Episodes

Eric Jang: AI is Good For You
Answers 383 questions
Kyunghyun Cho: Neural Machine Translation, Language, and Doing Good Science
Answers 383 questions
Jeffrey Ding on China's AI Dream, the AI 'Arms Race', and AI as a General Purpose Technology
Answers 383 questions
Greg Yang on Communicating Research, Tensor Programs, and µTransfer
Answers 383 questions
Hugo Larochelle: Deep Learning as Science
Answers 383 questions
Jeff Clune: Genetic Algorithms, Quality-Diversity, Curiosity
Answers 383 questions
Sergey Levine on Robot Learning & Offline RL
Answers 383 questions
Yoshua Bengio: The Past, Present, and Future of Deep Learning
Answers 383 questions
Zachary Lipton: Where Machine Learning Falls Short
Answers 383 questions
Joon Park: Generative Agents and Human-Computer Interaction
Answers 383 questions
Chelsea Finn on Meta Learning & Model Based Reinforcement Learning
Answers 383 questions
Max Braun: Teaching Robots to Help People in their Everyday Lives
Answers 383 questions
Sebastian Raschka: AI Education and Research
Answers 383 questions
Linus Lee: At the Boundary of Machine and Mind
Answers 383 questions
Yannic Kilcher on Being an AI Researcher and Educator
Answers 383 questions

Dexa/Daniel Bashir

Eric Jang on Robots Learning at Google and Generalization via Language

Topics covered

Popular Clips

Robotic Meta Learning

Unsupervised Object Learning

Meta Learning Robotics

Prompt Engineering Discovery

Reinforcement Learning Challenges

Generalization Challenges in Robotics

AI Generalization Insights

Future of Robotic Learning

Language and Generalization

AI Evolution

Language-Driven Task Learning

Language and AI

Language and Generalization

Robot Learning Challenges

Episode Highlights

Language and Robotics

Robotics Generalization

Semantic Grasping

Robotic Learning Techniques

Related Episodes

Eric Jang: AI is Good For You

Kyunghyun Cho: Neural Machine Translation, Language, and Doing Good Science

Jeffrey Ding on China's AI Dream, the AI 'Arms Race', and AI as a General Purpose Technology

Greg Yang on Communicating Research, Tensor Programs, and µTransfer

Hugo Larochelle: Deep Learning as Science

Jeff Clune: Genetic Algorithms, Quality-Diversity, Curiosity

Sergey Levine on Robot Learning & Offline RL

Yoshua Bengio: The Past, Present, and Future of Deep Learning

Zachary Lipton: Where Machine Learning Falls Short

Joon Park: Generative Agents and Human-Computer Interaction

Chelsea Finn on Meta Learning & Model Based Reinforcement Learning

Max Braun: Teaching Robots to Help People in their Everyday Lives

Sebastian Raschka: AI Education and Research

Linus Lee: At the Boundary of Machine and Mind

Yannic Kilcher on Being an AI Researcher and Educator

Eric Jang on Robots Learning at Google and Generalization via Language

Topics covered

Popular Clips

Episode Highlights

Language and RoboticsEric Jang discusses the transformative role of language in enhancing task learning and generalization for robotics. He highlights how language conditioning surpasses video-based methods, enabling robots to perform zero-shot tasks and understand complex concepts.

Language and Robotics

Robotics GeneralizationEric Jang explores the complexities of achieving task generalization in robotics, emphasizing the challenges of data collection and the potential of meta-learning. He highlights the need for robust benchmarks and innovative data methodologies to advance robotic capabilities.

Robotics Generalization

Semantic GraspingEric Jang, a research scientist at Google, discusses his pioneering work in semantic grasping for robotics. He highlights the evolution of techniques that enable robots to intelligently identify and manipulate objects, advancing their practical applications.

Semantic Grasping

Robotic Learning TechniquesEric Jang explores the integration of meta learning and reinforcement learning in robotics, highlighting advancements in task generalization and efficiency. He discusses the role of supervised learning in stabilizing robotic processes and enhancing capabilities.

Robotic Learning Techniques

Related Episodes