Skip to content

Rutwik99/Deep-Learning-ADMM-Optimizer

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ADMM as an Optimizer in Deep Learning

Goal

This project explores the application of the Alternating Direction Method of Multipliers (ADMM) as an optimizer for deep learning models, comparing its performance with established methods such as Stochastic Gradient Descent (SGD) and Adaptive Moment Estimation (ADAM).

Introduction

  • The Alternating Direction Method of Multipliers (ADMM) is a powerful optimization technique that finds its significance in various domains, making it essential to understand its core principles. ADMM combines elements of gradient descent and dual decomposition and is widely employed to solve complex optimization problems, particularly those with constraints. It offers a versatile framework for handling both equality and inequality constraints, making it valuable for a range of applications.
  • ADMM tries to combine the advantages of two methods, Dual Ascent and Augmented Lagrangian.
  • ADMM (Alternating Direction Method of Multipliers) divides the problem into multiple subproblems that can be tackled in parallel, eliminating the vanishing gradient issue without the need for gradient steps. This subdivision also brings greater stability to the optimization process.

Installation

To install the necessary dependencies, run:

pip install pytorch scikit-learn

To run the main script:

python main.py

Conclusion

The ADMM optimizer demonstrates its potential in deep learning by providing stability and efficient constraint handling. Although it might not always outperform Adam and Adagrad, it offers a robust alternative for specific applications requiring effective constraint management.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published