Binary Stochastic Filtering: a Method for Neural Network Size Minimization and Supervised Feature Selection

TL;DR

Binary Stochastic Filtering (BSF) is a novel method that effectively reduces neural network size and selects minimal feature subsets by integrating a stochastic filtering layer into the training process, improving efficiency and accuracy.

Contribution

The paper introduces BSF, a new stochastic filtering approach for neural network pruning and feature selection that outperforms existing methods in size reduction and accuracy.

Findings

Achieved significant neural network size reduction.

Surpassed literature in feature selection accuracy/dimensionality ratio.

Demonstrated effectiveness across multiple experiments.

Abstract

Binary Stochastic Filtering (BSF), the algorithm for feature selection and neuron pruning is proposed in this work. The method defines filtering layer which penalizes amount of the information involved in the training process. This information could be the input data or output of the previous layer, which directly leads to the feature selection or neuron pruning respectively, producing \textit{ad hoc} subset of features or selecting optimal number of neurons in each layer. Filtering layer stochastically passes or drops features based on individual weights, which are tuned with standard backpropagation algorithm during the training process. Multifold decrease of neural network size has been achieved in the experiments. Besides, the method was able to select minimal number of features, surpassing literature references by the accuracy/dimensionality ratio.