← Search

Soorya Gopalakrishnan

2 accepted papers

2020

Polarizing Front Ends for Robust Cnns

ICASSP 2020accepted

The vulnerability of deep neural networks to small, adversarially designed perturbations can be attributed to their "excessive linearity." In this paper, we propose a bottom-up strategy for attenuating adversarial perturbations using a nonlinear front end which polarizes and quantizes the data. We o…

Cited by 0SourceScholar
2017

Compressive information acquisition with hardware impairments and constraints: A case study

ICASSP 2017accepted

Compressive information acquisition is a natural approach for low-power hardware front ends, since most natural signals are sparse in some basis. Key design questions include the impact of hardware impairments (e.g., nonlinearities) and constraints (e.g., spatially localized computations) on the fid…

Cited by 0SourceScholar