Published May 19, 2020

ICLR 2020: Yann LeCun and Energy-Based Models

Tim Scarfe hosts a compelling episode featuring Yann LeCun at ICLR 2020, delving into the revolutionary concepts of self-supervised and contrastive learning, and the unifying power of energy-based models in transforming AI and machine learning paradigms.

Episode Highlights

Topics covered

Episode Highlights

Related Episodes

ICLR 2020: Yoshua Bengio and the Nature of Consciousness
Answers 383 questions
#86 - Prof. YANN LECUN and Dr. RANDALL BALESTRIERO - SSL, Data Augmentation, Reward isn't enough [NEURIPS2022]
Answers 383 questions
061: Interpolation, Extrapolation and Linearisation (Prof. Yann LeCun, Dr. Randall Balestriero)
Answers 383 questions
#032- Simon Kornblith / GoogleAI - SimCLR and Paper Haul!
Answers 383 questions
Francois Chollet - On the Measure of Intelligence
Answers 383 questions
#55 Self-Supervised Vision Models (Dr. Ishan Misra - FAIR).
Answers 383 questions
Dr. Paul Lessard - Categorical/Structured Deep Learning
Answers 383 questions
#51 Francois Chollet - Intelligence and Generalisation
Answers 383 questions
#102 - Prof. MICHAEL LEVIN, Prof. IRINA RISH - Emergence, Intelligence, Transhumanism
Answers 383 questions
Can we build a generalist agent? Dr. Minqi Jiang and Dr. Marc Rigter
Answers 383 questions
Mahault Albarracin - Cognitive Science
Answers 383 questions
CURL: Contrastive Unsupervised Representations for Reinforcement Learning
Answers 383 questions
#036 - Max Welling: Quantum, Manifolds & Symmetries in ML
Answers 383 questions
Pattern Recognition vs True Intelligence - Francois Chollet
Answers 383 questions
Jürgen Schmidhuber - Neural and Non-Neural AI, Reasoning, Transformers, and LSTMs
Answers 383 questions

Dexa/Machine Learning Street Talk (MLST)

ICLR 2020: Yann LeCun and Energy-Based Models

Topics covered

Popular Clips

Energy Functions in ML

Probabilistic vs. Energy Models

Energy Based Models

Bert's Dual Objectives

Neural Network Manifolds

Neural Network Predictions

General Intelligence Debate

Generative Models Discussion

Energy Functions vs. Probabilistic Models

Model-Based Reinforcement Learning

Language Manifold Exploration

GANs and Representation

Understanding Data Manifolds

Energy-Based Learning

Multiple Video Frames

Episode Highlights

Self-Supervised Learning

Contrastive Learning

Energy-Based Models

Related Episodes

ICLR 2020: Yoshua Bengio and the Nature of Consciousness

#86 - Prof. YANN LECUN and Dr. RANDALL BALESTRIERO - SSL, Data Augmentation, Reward isn't enough [NEURIPS2022]

061: Interpolation, Extrapolation and Linearisation (Prof. Yann LeCun, Dr. Randall Balestriero)

#032- Simon Kornblith / GoogleAI - SimCLR and Paper Haul!

Francois Chollet - On the Measure of Intelligence

#55 Self-Supervised Vision Models (Dr. Ishan Misra - FAIR).

Dr. Paul Lessard - Categorical/Structured Deep Learning

#51 Francois Chollet - Intelligence and Generalisation

#102 - Prof. MICHAEL LEVIN, Prof. IRINA RISH - Emergence, Intelligence, Transhumanism

Can we build a generalist agent? Dr. Minqi Jiang and Dr. Marc Rigter

Mahault Albarracin - Cognitive Science

CURL: Contrastive Unsupervised Representations for Reinforcement Learning

#036 - Max Welling: Quantum, Manifolds & Symmetries in ML

Pattern Recognition vs True Intelligence - Francois Chollet

Jürgen Schmidhuber - Neural and Non-Neural AI, Reasoning, Transformers, and LSTMs

ICLR 2020: Yann LeCun and Energy-Based Models

Topics covered

Popular Clips

Episode Highlights

Self-Supervised LearningThe exploration of self-supervised learning reveals its potential to transform AI by enabling machines to learn from unlabeled data. This approach, championed by Yann LeCun, offers a new paradigm for developing intelligent systems.

Self-Supervised Learning

Contrastive LearningThe discussion shifts to understanding contrastive learning methods and their significance in machine learning. These methods, crucial for representation learning, are explored through their applications in GANs and the challenges of negative mining strategies.

Contrastive Learning

Energy-Based Models

Related Episodes