-
A 15-minute video description of Sourya’s research on exploring complexity reduction in deep learning
-
Pre-defined sparsity for reducing complexity in neural networks - Sourya Dey
An article explaining some of the work in our paper ‘Pre-defined Sparse Neural Networks with Hardware Acceleration’. - Sparse Matrices in Pytorch - Sourya Dey
A series of articles analyzing runtimes of sparse matrices and their dense counterparts in Pytorch. -
Morse code dataset for Artificial Neural Networks - Sourya Dey
This article explains some part of the work in our paper ‘Morse Code Datasets for Machine Learning’. - Neural Networks Training with Approximate Logarithmic Computations - Arnab Sanyal
An article explaining some of the work in our paper with the same title.