Unsupervised Pre-Training by Evolving for Diverse Features

Presenter's Name(s)

Jordan Donovan

Conference Year

January 2023

Abstract

Deep neural networks (DNNs) excel at extracting complex patterns from data to solve complex, non-linear problems across several domains. The various initialization strategies utilized can greatly affect the accuracy of the resulting trained network and efficiency of the training process. We propose an evolutionary pre-training technique that initializes networks in a manner that optimizes toward orthogonality of feature activations in the convolutional filters of a CNN. Relative to randomly initialized parameters, this evolutionary pre-training improves the resulting accuracy of networks when training these convolutional filters on the image classification benchmark CIFAR-100.

Primary Faculty Mentor Name

Nicholas Cheney

Status

Graduate

Student College

College of Engineering and Mathematical Sciences

Program/Major

Computer Science

Primary Research Category

Engineering and Math Science

Abstract only.

Share

COinS
 

Unsupervised Pre-Training by Evolving for Diverse Features

Deep neural networks (DNNs) excel at extracting complex patterns from data to solve complex, non-linear problems across several domains. The various initialization strategies utilized can greatly affect the accuracy of the resulting trained network and efficiency of the training process. We propose an evolutionary pre-training technique that initializes networks in a manner that optimizes toward orthogonality of feature activations in the convolutional filters of a CNN. Relative to randomly initialized parameters, this evolutionary pre-training improves the resulting accuracy of networks when training these convolutional filters on the image classification benchmark CIFAR-100.