Efficient Deep Feature Learning and Extraction via StochasticNets

Abstract

Deep neural networks are a powerful tool for feature learning and extraction. One area worth exploring in feature extraction using deep neural networks is efficient neural connectivity formation for faster feature learning and extraction. Motivated by findings of stochastic synaptic connectivity formation in the brain as well as the brain's uncanny ability to efficiently represent information, we propose the efficient learning and extraction of features via StochasticNets, where sparsely-connected deep neural networks can be formed via stochastic connectivity between neurons. Experimental results show that features learned using deep convolutional StochasticNets, with fewer neural connections than conventional deep convolutional neural networks, can allow for better or comparable classification accuracy than conventional deep neural networks. Finally, it was also shown that significant gains in feature extraction speed can be achieved in embedded applications using StochasticNets.

Related Papers

• → Conditional Deep Learning for Energy-Efficient and Enhanced Pattern Recognition(2016)141 cited
• → CSAERNet: An Efficient Deep Learning Architecture for Image Classification(2020)21 cited
• → Parametric Learning of Deep Convolutional Neural Network(2014)2 cited
• → A Hybrid Model for Classification of Biomedical Data Using Feature Filtering and a Convolutional Neural Network(2018)2 cited