Faster gradient descent via an adaptive learning rate

This repo contains code for work done by Mathieu Ravaut and Satya Krishna Gorti.

In this project we show a new adaptive learning rate algorithm on gradient descent and explore its effectiveness on the following tasks:

• Linear regression using Boston Housing Prices dataset
• Logistic regression using MNIST dataset
• Image Classification with neural networks using CIFAR-10 and CIFAR-100

The report of this work can be seen here