Distinction between features extracted using deep belief networks

TL;DR

This paper proposes two methods to distinguish relevant features from irrelevant ones extracted by Deep Belief Networks in face recognition, enhancing understanding of feature importance for machine learning tasks.

Contribution

The paper introduces two novel methods for distinguishing task-relevant features from irrelevant features in features extracted by Deep Belief Networks.

Findings

Methods successfully differentiate relevant features in face recognition.

Application of methods improves interpretability of deep features.

Demonstrates importance of feature relevance in machine learning performance.

Abstract

Data representation is an important pre-processing step in many machine learning algorithms. There are a number of methods used for this task such as Deep Belief Networks (DBNs) and Discrete Fourier Transforms (DFTs). Since some of the features extracted using automated feature extraction methods may not always be related to a specific machine learning task, in this paper we propose two methods in order to make a distinction between extracted features based on their relevancy to the task. We applied these two methods to a Deep Belief Network trained for a face recognition task.