Deep Learning of Generative Models - 2014

Explore the cutting-edge field of deep learning and generative models in this comprehensive lecture by renowned AI researcher Yoshua Bengio. Delve into key concepts such as unsupervised learning, object recognition, and language processing. Discover how these techniques are revolutionizing areas like computer vision, speech recognition, and image search. Examine computational bottlenecks, sampling methods, and the importance of good representations in AI systems. Learn about innovative approaches like analogy learning, the free trading trick, and dependency nets. Investigate fundamental problems in the field and their potential solutions, including experiments with default machines. Gain insights into the future of AI and its applications across various domains in this informative 71-minute talk from the Center for Language & Speech Processing at Johns Hopkins University.

Intro
Motivations
Google Image Search
Language Processing
Analogy Learning
Computational bottleneck
Sampling methods
Object recognition
MIT Technology Review 2013
Speech Recognition
Computer Vision
Unsupervised Learning
Free Trading Trick
What is a good representation
Priors
Improvised Learning
Fundamental Problems
First Problem
Experiments
Dependency Nets
Default Machine
Default Machine Results
Noise