Theory of Generative Deep Learning : Probe Landscape of Empirical Error via Norm Based Capacity Control

TL;DR

This paper explores the theoretical landscape of generative deep learning by analyzing empirical error and capacity control, providing insights into training dynamics through mathematical and biological perspectives.

Contribution

It introduces a novel theoretical framework for understanding generative deep learning, focusing on empirical error landscape and norm-based capacity control.

Findings

Highlights the empirical error landscape in generative models

Provides a mathematical interpretation of training dynamics

Connects biological insights with deep learning theory

Abstract

Despite its remarkable empirical success as a highly competitive branch of artificial intelligence, deep learning is often blamed for its widely known low interpretation and lack of firm and rigorous mathematical foundation. However, most theoretical endeavor is devoted in discriminative deep learning case, whose complementary part is generative deep learning. To the best of our knowledge, we firstly highlight landscape of empirical error in generative case to complete the full picture through exquisite design of image super resolution under norm based capacity control. Our theoretical advance in interpretation of the training dynamic is achieved from both mathematical and biological sides.