Shamak Dutta

s7dutta at edu dot uwaterloo dot ca

I am a graduate student at the University of Waterloo, where I work on machine learning and its intersection with computational neuroscience. I am part of the BRAIN Lab and the Machine Learning Research Group, advised by Bryan Tripp and Graham Taylor.

I am currently interning at Preferred Networks in Tokyo, Japan. I'll be working on problems related to computer vision.

Before graduate school, I received a Bachelors in ECE at the University of Waterloo, where I worked on several research projects, including behaviour cloning for human motion under Dana Kulic , the generalized travelling salesman problem under Stephen Smith and compressed image retrieval under H.R. Tizhoosh. I have also completed seven internships across six different companies including Amazon Search, Amazon Advertising and Morpheus Labs.

CV  /  Google Scholar  /  GitHub

Research

I am interested in understanding the similarities and differences between information processing in the visual cortex and convolutional neural networks. I am also interested in optimization and its use for non-differentiable and stochastic functions.

Convolutional Neural Networks Regularized by Correlated Noise
Shamak Dutta, Bryan Tripp, Graham Taylor
Conference on Computer and Robot Vision, 2018
bibtex

We show that adding spatially and tuning dependent correlated noise to the activities of units in a neural network helps in regularization and image classification under occlusion.

Barcodes for Medical Image Retrieval Using Autoencoded Radon Transform
H.R. Tizhoosh, Christopher Mitcheltree, Shujin Zu, Shamak Dutta
International Conference on Pattern Recognition, 2016
bibtex

We show that barcodes generated by thresholding the learned representations of autoencoders on the Radon transform of images can be used in medical image search and retrieval.


credits