Arya Mazumdar (Publications)

Copyright: Most of the papers listed below are published and the copyrights may have been transferred to the publisher. Papers cannot be duplicated for commercial purposes.

Author ordering: I try to maintain alphabetical ordering of authors in my papers. However, as per engineering tradition student/junior authors are sometime listed first.

J: Journal. Unmarked: Conferences.

Preprints

2018

  • Capacity of Locally Recoverable Codes
    Arya Mazumdar
    IEEE Information Theory Workshop (ITW), 2018.

  • Novel Impossibility Results for Group-Testing
    Abhishek Agarwal, Sidharth Jaggi, Arya Mazumdar
    IEEE International Symposium on Information Theory, 2018.

  • (J18)Combinatorial Alphabet-Dependent Bounds for Locally Recoverable Codes
    Abhishek Agarwal, Alexander Barg, Sihuang Hu, Arya Mazumdar, Itzhak Tamo
    IEEE Transactions on Information Theory, vol. 64, no. 5, May 2018.

  • Robust Gradient Descent via Moment Encoding and LDPC codes (Extended Abstract)
    Raj Kumar Maity, Ankit Singh Rawat, Arya Mazumdar
    The SysML Conference, 2018.
    One of the 7 Oral presentations out of 205 submissions.

  • The Geometric Block Model
    Sainyam Galhotra, Arya Mazumdar, Soumyabrata Pal, Barna Saha
    The Thirty-Second AAAI Conference on Artificial Intelligence (AAAI-18).
    The AAAI proceeding version contains some errors that have since been corrected in the above revised version in arXiv. Namely, in the Geometric Block Model, given two vertices connected by an edge, the events of other vertices being a common neighbor of both are not independent events (unlike the stochastic block model). Please refer to the arXiv version linked above for corrected statements and results.

2017

2016

2015

2014

2013

2012

2011

2010

2009 and earlier

Thesis

  • Combinatorial Methods in Coding Theory
    Ph.D. Thesis, University of Maryland, 2011. Distinguished Dissertation Award.
    My work in the thesis is essentially equal to the sum of contents of journal publications marked with a #.