Vignesh Subramanian

vignesh.subramanian[AT]berkeley[DOT]edu

PhD Student
Electrical Engineering and Computer Science
University of California Berkeley

I am a fifth year PhD student in the EECS department at UC Berkeley, advised by Prof. Anant Sahai.

My research interests lie primarily in the theoretical understanding of machine learning and its application to wireless communication and control. More recently I have become interested in using machine learning for object detection and tracking in autonomous driving systems.

Prior to joining UC Berkeley, I worked for two years at WorldQuant Research, India in Mumbai as Senior Quantitative Researcher. During the summer of 2021, I did an internship at Plus, an autonomous trucking techonology company.

I graduated from Indian Institute of Techonology, Bombay with a B.Tech + M.Tech (Dual degree) in Electrical Engineering and Minors in Computer Science and Engineering . I was awarded the Institue Gold Medal and Institute Silver Medal during my graduation. I was fortunate to have worked with Prof. Sibi Raj Pillai and Prof. V. Rajbabu on my Master's thesis.
:)

Current & Past Affiliations

University of California Berkeley | Fourth year PhD student, Electrical Engineering and Computer Science
Advisor: Prof Anant Sahai | Aug 2017-Present
Plus, Cupertino | Machine Learning Internship | May-August 2021
Worldquant Research, India | Quantitative Researcher | July 2015-July 2017
Indian Institute of Technology Bombay | Bachelor of Technology & Master of Technology, Electrical Engineering
Advisors: Prof Sibiraj Pillai & Prof V Rajbabu| 2010-2015
Bell Labs Alcatel Lucent, Bangalore | Research Internship
Advisor: Anand Muralidhar | May-July 2013

My Work

For a PDF version, please see here: [short], [long].

Machine Learning for Perception| Internship (May - Aug 2021)
Plus, Cupertino

Worked on state of the art image based anchor-free object detection and tracking implementation in PyTorch.

Machine learning for Physical Layer Wireless Communication | Aug 2018 - Present
UC Berkeley | Collaborators: Josh Sanz, Caryn Tran, Kailas Vodrahalli, Prof. Anant Sahai

Designed a blind interactive learning protocol for modulation schemes in the multi-agent setting without codesign. Experimentally verified the universality and robustness of the protocol and showed that it achieves bit error rates similar to the optimal baseline. Working on integrating other parts of the communication pipeline including equalization and error correcting codes to enable end-to-end learning of communication schemes. 

Learning Stabilizing Control under Multiplicative Noise | July 2019 - January 2020
UC Berkeley| Collaborators: Moses Won, Prof. Gireeja Ranade

Exploring use of neural networks to discover control strategies for stabilizing a system under multiplicative noise. Proposed an architecture and training procedure tailored for the control problem that enables the network to generalize and output controls for rollouts longer than the training horizon. Showed that the neural network based control strategybeats current best known strategies including optimal linear strategies. 

Classification versus Regression for Minimum Norm Interpolating Solutions | August 2019 - Present
UC Berkeley | Collaborators: Vidya Muthukumar, Adhyyan Narang, Prof Anant Sahai; UC Berkeley, Prof. Daniel Hsu; Columbia University, Prof. Mikhail Belkin; Ohio State University

Analyzed the classification and regression loss of minimum norm interpolating solutions in the overparameterized setting. Related the classification error to statistical signal processing concepts of shrinkage and false-discovery and computed sharp upper and lower bounds for these quantities. Showed the existence of a regime where asymptotically classification performs well but regression does not.

Neural Network based Control for Witsenhausen problem | Jan-Aug 2018
UC Berkeley | Guide: Prof. Anant Sahai

Revisited the classical decentralized stochastic control problem of Witsenhausen using neural networks. Concluded that biasing architectures towards favorable solutions is required to escape the local minimas of the non-convex problem. Numerically showed that in higher dimensions it is possible to outperform the best known one dimensional strategy. 

Compressed Sensing in Radio Astronomy | Master's Thesis (July 2014 - May 2015)
Indian Institute of Technology Bombay | Advisors: Prof. Sibi Raj Pillai & Prof. V. Rajbabu 

Thesis PDF : here

Analyzed iterative shrinkage-thresholding algorithms for recovering images of astronomical sources from a set of highly incomplete Fourier measurements. Formulated the simultaneous reconstruction of multiple images as a joint minimization problem and proposed an alternating algorithm to solve the problem. Tested the alternating algorithm on images that are sparse in either spatial or wavelet domain and concluded that simultaneous reconstruction results in better performance as compared to independent reconstructions.

Scheduling Algorithms for Wireless Communication | Research Internship (May - July 2013)
Bell Labs Alcatel Lucent, Bangalore | Advisor: Anand Muralidhar

Analyzed the performance of scheduling algorithms for device to device communication in cellular networks. Implemented and tested a novel coloring algorithm in Python and concluded that it increases total network throughput by up to 70% as compared to existing schemes like CSMA/CA. Theoretically proved approximate optimality of the greedy coloring algorithm by obtaining bounds on its performance.

Publications

  1. Vidya Muthukumar, Adhyyan Narang, Vignesh Subramanian, Mikhail Belkin, Daniel Hsu, Anant Sahai "Classification vs regression in overparameterized regimes: Does the loss function matter? ", Journal of Machine Learning Research (JMLR), 2020. [link]
     
  2. Vignesh Subramanian, Moses Won, Gireeja Ranade, "Learning a Neural-Network Controller for a Multiplicative Observation Noise System ", IEEE International Symposium on Information Theory (ISIT) , 2020. [link]
     
  3. Anant Sahai, Joshua Sanz, Vignesh Subramanian, Caryn Tran, Kailas Vodrahalli, "Blind interactive learning of modulation schemes: Multi-agent cooperation without co-design ", IEEE Access, Special Section: Artificial Intelligence for Physical-layer Wireless, 2019. [link]
     
  4. Vidya Muthukumar, Kailas Vodrahalli, Vignesh Subramanian, Anant Sahai, "Harmless interpolation of noisy data in regression", IEEE Journal on Selected Areas in Information Theory, Special Issue on Deep Learning: Mathematical Foundations and Applications to Information Science, 2019. [link]
     
  5. Anant Sahai, Joshua Sanz, Vignesh Subramanian, Caryn Tran, Kailas Vodrahalli,"Learning to communicate with limited co-design", 57th Annual Allerton Conference on Communication, Control, and Computing, 2019. [link]
     
  6. Vignesh Subramanian, Laura Brink, Nikunj Jain, Kailas Vodrahalli, Akhil Jalan, Nikhil Shinde, Anant Sahai," Some new numeric results concerning the Witsenhausen Counterexample", 56th Annual Allerton Conference on Communication, Control, and Computing, 2018 [link]

Teaching

Graduate Student Instructor | EECS 189/289A, Machine Learning | Fall 2020
Graduate Student Instructor | EECS 127/227A, Optimization | Spring 2019, Spring 2020

Contact

Email:   vignesh.subramanian[AT]berkeley[DOT]edu

Office:   BLISS, 264 Cory Hall, Berkeley (Google Maps)

LinkedIn Profile

Google Scholar Profile