Cadambe-Viveck



Contact


Email: viveck@psu.edu
Room: 230, Electrical Engineering West,
Pennsylvania State University,
University Park, PA 16802


I am an Associate Professor in the Department of Electrical Engineering at Pennsylvania State University. I received my Ph.D from the University of California, Irvine in 2011. Between 2011 and 2014, I was a postdoctoral researcher, jointly with the Electrical and Computer Engineering department at Boston University, and the Research Laboratory of Electronics (RLE) at the Massachusetts Institute of Technology (MIT). I received the 2009 IEEE Information Theory Society Best Paper Award, the 2014 IEEE International Symposium on Network Computing and Applications (NCA) Best Paper Award, an NSF Career Award in 2016, was a finalist for the Bell Labs Prize in 2016, and a 2019 Google Faculty Award. I was an Associate Editor for IEEE Transactions on Wireless Communications between 2014 and 2019. In my research, I study mathematical foundations of learning, computing and communications using tools from information and coding theory, distributed systems theory and optimization.

News:
  • Paper titled epsilon-Approximate Coded Matrix Multiplication is Nearly Twice as Efficient as Exact Multiplication with authors Haewon Jeong, Ateet Devulapalli, Viveck R. Cadambe and Flavio Calmon, accepted to IEEE ISIT 2021, extended version to appear in 2021 Journal of Special Topics in Information Theory (JSAIT), 2021, draft on arxiv:2105.01973
  • Paper titled Extreme Point Approach to Subset Selection with authors Srikanth Banagere, Viveck R. Cadambe and Bill Kay, a spotlight presentation at SubSetML: Subset Selection in Machine Learning: From Theory to Practice Workshop at 2021 ICML Conference
  • Paper titled Local SGD with Periodic Averaging: Tighter Analysis and Adaptive Synchronization with authors Farzin Haddadpour, Mohammad Mahdi Khamani, Mehrdad Mahdavi and Viveck R. Cadambe accepted to 2019 Conference on Neural Information Processing Systems (NeurIPS).
  • Farzin Haddadpour his thesis titled "Communication-Efficient and Fault-Tolerant Algorithms for Distributed Machine Learning", congratulations!!
  • Tutorial and panel on coded computing in ISIT 2020 along with Salman Avestimehr and Suhas Diggavi see video here
  • Ramy Ali defended his thesis titled "Harnessing Data Correlation and Network Information in Distributed Key-Value Stores", congratulations!!
  • Mohammad Fahim defended his thesis titled "Polynomial Approaches for Network Coding and Fault-Tolerant Distributed Computing", congratulations!!
  • Teaching a Special Topics course on Distributed Computing in Spring 2019.
  • Paper titled Local SGD with Periodic Averaging: Tighter Analysis and Adaptive Synchronization with authors Farzin Haddadpour, Mohammad Mahdi Khamani, Mehrdad Mahdavi and Viveck R. Cadambe accepted to 2019 Conference on Neural Information Processing Systems (NeurIPS).
  • Paper titled Lagrange Coded Computing with Sparsity Constraints with authors Mohammad Fahim and Viveck R. Cadambe presented to 2019 Allerton Conference on Communications, Control and Computing.
  • Co-organized 2019 Coding Theory for Large-Scale Machine Learning (CodML) Workshop, co-located with ICML 2019.
  • Paper titled Trading Redundancy for Communication: Speeding up Distributed SGD for Non-Convex Optimization with authors Farzin Haddadpour, Mohammad Mahdi Kamani, Mehrdad Mahdavi, Viveck R. Cadambe accepted to International Conference on Machine Learning (ICML 2019).
  • Paper titled ARES: Adaptive, Reconfigurable, Erasure coded, Distributed Storage , with authors Nicolas Nicolau, Viveck R. Cadambe, Kishori Konwar, N. Prakash, Muriel Medard and Nancy Lynch, accepted to 2019 IEEE International Conference on Distributed Computing Systems (ICDCS) available on arXiv.
  • Talk titled "How to make polynomially coded computing numerically stable'' at ITA Workshop in Feb 2019, presents joint work with Mohammad T. Fahim. Extended paper on arXiv
  • Paper titled Cross-Iteration Coded Computing with authors Farzin Haddadpour, Yaoqing Yang, Viveck R. Cadambe and Pulkit Grover presented by Farzin Haddadpour presented at 2018 Annual Allerton Conference on Communication, Control and Computing
  • Talk titled "Error Correcting Codes for Distributed Algorithms and Computing" presented at Oak Ridge National Lab, July 2018
  • Paper titled Codes for Distributed Finite Alphabet Matrix-Vector Multiplication with authors Farzin Haddadpour and Viveck R. Cadambe presented by Farzin Haddadpour presented at 2018 IEEE International Symposium on Information Theory.
  • Paper titled Multi-Version Coding with Side-Information with authors Ramy E. Ali, Viveck R. Cadambe, Jaime Llorca and Antonia Tulino presented by Ramy Ali at 2018 IEEE International Symposium on Information Theory.
  • Paper titled Straggler-Resilient and Communication Efficient Distributed Iterative Linear Solver , with authors Farzin Haddadpour, Yaoqing Yang, Malhar Chaudhari, Viveck R. Cadambe, and Pulkit Grover available on arXiv.
  • Paper titled On the Optimal Recovery Threshold of Coded Matrix Multiplication , with authors Sanghamitra Dutta, Mohammad Fahim, Farzin Haddadpour, Haewon Jeong, Viveck R. Cadambe, Pulkit Grover available on arXiv. Extended version of previously published paper at Allerton 2017.
  • Paper titled Multi-version Coding for Consistent Distributed Storage of Correlated Data Updates, with authors Ramy E. Ali and Viveck R. Cadambe available on arXiv. Extended version of previously published paper at ITW 2016.
  • Paper titled Coded convolution for parallel and distributed computing within a deadline with authors Sanghamittra Dutta, Viveck R. Cadambe and Pulkit Grover presented in the 2017 IEEE International Symposium on Information Theory (ISIT). Extended version available here.
  • Paper titled Linear Network Coding for Two-Unicast-Z Networks: Commutative-Algebraic Perspective and Fundamental Limits with authors Mohammad Fahim and Viveck R. Cadambe presented in the 2017 IEEE International Symposium on Information Theory (ISIT). Extended version available here.
  • Paper titled Short-Dot: Computing Large Linear Transforms Distributedly Using Coded Short Dot Products, with authors Sanghamittra Dutta, Viveck R. Cadambe and Pulkit Grover, in The Thirtieth Annual Conference on Neural Information Processing Systems (NIPS), 2016.
  • Delivered invited talk on "Asynchrony-Storage Tradeoffs in Consistent Distributed Storage Systems" at Asilomar Conference on Signals and Systems 2016. Slides available here.
  • Paper titled Consistent Distributed Storage of Correlated Updates Via Multi-version Coding, with authors Ramy Ali and Viveck Cadambe in 2016 Information Theory Workshop.
  • Paper titled Information-Theoretic Lower Bounds on the Storage Cost of Shared Memory Emulation, with authors Viveck Cadambe, Zhiying Wang and Nancy Lynch in 2016 Principles of Distributed Computing (PODC) conference. Extended version of the paper can be found here.
  • Paper titled Alphabet-size Dependent Bounds for Exact Repair in Distributed Storage", Viveck Cadambe and Arya Mazumdar, at 2015 IEEE Information Theory Workshop (ITW), Jeju Island.
  • Paper titled "Multi-version Coding - An Information Theoretic Perspective of Distributed Storage" with authors Zhiying Wang, Viveck Cadambe available on arxiv; extended version of our papers that appeared in Proceedings of IEEE International Symposium on Information Theory (ISIT) 2014 and the 2014 Allerton Conference on Communications, Control and Computing
  • Paper titled "Expanding the Compute-and-Forward Framework: Unequal Powers, Signal Levels, and Multiple Linear Combinations" with authors Bobak Nazer, Viveck Cadambe, Vasilis Ntranos, Giuseppe Caire, available on arxiv
  • Talk titled "Multi-version Coding and Shared Memory Emulation" presented at Workshop on Coding: From Practice to Theory at the Simons Institute for Theory of Computing, Berkeley, Feb 2015.
  • Talk titled "Alignment Based Network Coding For Two-Unicast-Z Networks" presented at Information Theory and Applications (ITA) Workshop, Feb 2015. See extended version of paper here.