Главная
Study mode:
on
1
Intro
2
Why Deep Networks?
3
Data Models and Deep Networks
4
The Dream
5
The Pure Theorist Model
6
': A DL Theorist Perspective
7
Hacker models
8
Candidate 3: Scattering Transform
9
The Question Remains
10
Information Flow on Trees
11
Is this process natural?
12
What is the best classifier
13
Provable Algorithms for learning classifier
14
Depth Lower bounds
15
Phylogenetic Reconstruction
16
A semi supervised setting
17
Deep Algorithms
Description:
Explore the fundamentals and applications of generative models in this lecture by Elchanan Mossel from the Massachusetts Institute of Technology, presented at the Deep Learning Boot Camp. Delve into the rationale behind deep networks, data models, and the theoretical perspectives of deep learning. Examine various models, including the Pure Theorist Model and Hacker models, and investigate the Scattering Transform. Analyze information flow on trees, natural processes, and optimal classifiers. Learn about provable algorithms for learning classifiers, depth lower bounds, and phylogenetic reconstruction. Gain insights into semi-supervised settings and deep algorithms in this comprehensive exploration of generative models and their implications in deep learning.

Generative Models

Simons Institute
Add to list