Selected Publications

Recently, there has been a wealth of effort devoted to the design of protocols for secure machine learning algorithms. In particular, much of this is aimed at ensuring predictions from highly-accurate deep neural network (NN) models are secure. However, as NNs are trained on data, a key question is how such models can be trained securely. The few prior works on secure NN training have focused either on designing custom protocols for existing training algorithms, or on developing tailored training algorithms and then applying generic secure protocols. In this work, we propose to simultaneously design training algorithms alongside a secure protocol for computing that algorithm, incorporating optimizations on both fronts. We present QUOTIENT, a new method for discretized training of deep neural networks designed to be evaluated in secure computation, along with a secure two-party protocol for it. QUOTIENT incorporates important components of state-of-the-art neural network training such as layer normalization and adaptive gradients. Compared to the state-of-the-art in secure two-party (2PC) neural network training, we obtain an improvement of 50X in time and 6% in accuracy. Additionally, our method is the first practical secure 2PC framework for neural network training over WAN.
To appear at ACM Conference on Computer and Communications Security (CCS’19), 2019

Recent Publications

. QUOTIENT : Two-Party Secure Neural Network Training and Predicition. To appear at ACM Conference on Computer and Communications Security (CCS’19), 2019.

. Iris recognition based on sparse representation and k-nearest subspace with genetic algorithm. Pattern Recognition Letters, 73, 13-18, Elsevier, 2016.

PDF

. Fast & Dynamic Image Restoration using Laplace equation Based Image Inpainting. DU Journal of undergraduate research and Innovation 1(1), University of Delhi, 2015.

PDF

. Parallel multi-objective multi-robot coalition formation. Expert Systems with Applications, 42(21), 7797-7811, Elsevier, 2015.

PDF

. GPU Accelerated B-spline Coefficient Computation. GPU Technology Conference, San Jose, CA, USA, 2015.

PDF

. A Graph Based Ranking Strategy for Automated Text Summarization. DU Journal of undergraduate research and Innovation 1(1), University of Delhi, 2015.

PDF

. Parallel multi-objective multi-robot coalition formation (Poster). GPU Technology Conference, San Jose, CA, USA, 2014.

PDF

. Chloroplast genes as genetic markers for inferring patterns of change, maternal ancestry and phylogenetic relationships among Eleusine species. AoB Plants, 6, Oxford University Press, 2014.

PDF

Awards & Competitive Funding

Experience

 
 
 
 
 
June 2018 – September 2018
London, UK

Research Intern (Privacy Preserving Machine Learning)

The Alan Turing Institute

 
 
 
 
 
March 2016 – June 2016
Delhi, India

Research Intern (Big Data Analytics)

IBM India Research Lab

 
 
 
 
 
June 2015 – August 2015
Nova Scotia, Canada

Mitacs Globalink Research Intern

Dalhousie University

 
 
 
 
 
June 2014 – August 2014
Mumbai, India

Research Intern (GPGPU Computing)

Indian Institute of Technology Bombay

 
 
 
 
 
June 2013 – May 2016
Delhi, India

Student Researcher (Cryptography & Machine Learning)

Defence Research & Development Organization

Recent & Upcoming Talks

Engagement

Contact