DeepAI AI Chat
Log In Sign Up

Training Integrable Parameterizations of Deep Neural Networks in the Infinite-Width Limit

by   Karl Hajjar, et al.

To theoretically understand the behavior of trained deep neural networks, it is necessary to study the dynamics induced by gradient methods from a random initialization. However, the nonlinear and compositional structure of these models make these dynamics difficult to analyze. To overcome these challenges, large-width asymptotics have recently emerged as a fruitful viewpoint and led to practical insights on real-world deep networks. For two-layer neural networks, it has been understood via these asymptotics that the nature of the trained model radically changes depending on the scale of the initial random weights, ranging from a kernel regime (for large initial variance) to a feature learning regime (for small initial variance). For deeper networks more regimes are possible, and in this paper we study in detail a specific choice of "small" initialization corresponding to ”mean-field” limits of neural networks, which we call integrable parameterizations (IPs). First, we show that under standard i.i.d. zero-mean initialization, integrable parameterizations of neural networks with more than four layers start at a stationary point in the infinite-width limit and no learning occurs. We then propose various methods to avoid this trivial behavior and analyze in detail the resulting dynamics. In particular, one of these methods consists in using large initial learning rates, and we show that it is equivalent to a modification of the recently proposed maximal update parameterization μP. We confirm our results with numerical experiments on image classification tasks, which additionally show a strong difference in behavior between various choices of activation functions that is not yet captured by theory.


Dynamics of Finite Width Kernel and Prediction Fluctuations in Mean Field Neural Networks

We analyze the dynamics of finite width effects in wide but finite featu...

Disentangling feature and lazy learning in deep neural networks: an empirical study

Two distinct limits for deep learning as the net width h→∞ have been pro...

On the Effect of Initialization: The Scaling Path of 2-Layer Neural Networks

In supervised learning, the regularization path is sometimes used as a c...

Initial Guessing Bias: How Untrained Networks Favor Some Classes

The initial state of neural networks plays a central role in conditionin...

Dynamically Stable Infinite-Width Limits of Neural Classifiers

Recent research has been focused on two different approaches to studying...

How to Start Training: The Effect of Initialization and Architecture

We investigate the effects of initialization and architecture on the sta...

Embedded Ensembles: Infinite Width Limit and Operating Regimes

A memory efficient approach to ensembling neural networks is to share mo...