A Path Towards Autonomous Machine Intelligence - Paper Explained
Offered By: Yannic Kilcher via YouTube
Course Description
Overview
Explore a comprehensive analysis of Yann LeCun's position paper on autonomous machine intelligence in this detailed video explanation. Delve into the integration of Self-Supervised Learning, Energy-Based Models, and hierarchical predictive embedding models to create a system capable of learning useful abstractions at multiple levels and utilizing them for future planning. Examine key concepts such as Mode 1 and Mode 2 actors, latent variables, the problem of collapse, and contrastive vs. regularized methods. Gain insights into the JEPA architecture and its hierarchical variant, H-JEPA. Understand the broader relevance of these concepts in the field of artificial intelligence and machine learning. Benefit from a thorough summary and expert commentary on this groundbreaking approach to developing autonomous intelligent agents.
Syllabus
- Introduction
- Main Contributions
- Mode 1 and Mode 2 actors
- Self-Supervised Learning and Energy-Based Models
- Introducing latent variables
- The problem of collapse
- Contrastive vs regularized methods
- The JEPA architecture
- Hierarchical JEPA H-JEPA
- Broader relevance
- Summary & Comments
Taught by
Yannic Kilcher
Related Courses
Artificial Intelligence Foundations: Thinking MachinesLinkedIn Learning Deep Learning for Computer Vision
NPTEL via YouTube NYU Deep Learning
YouTube Stanford Seminar - Representation Learning for Autonomous Robots, Anima Anandkumar
Stanford University via YouTube Towards Extreme Multi-Task Scaling for Transfer Learning
Yannic Kilcher via YouTube