Projects and Groups
- Computer Science
- Algorithms and Theory
- Artificial Intelligence
- Communications & Networking
- Knowledge Discovery and Data Mining
- Operations Research
- Signal Processing
- Smarter Energy
For my personal research page (with links to pdfs of my papers) go here .
Balakrishnan Narayanaswamy received his PhD from the department of Electrical and Computer Engineering (ECE) at Carnegie Mellon University in 2011, after which he joined the IBM Research Lab in Bangalore, India.
He graduated by successfully defending this :
His research interests lie at the intersection of AI, optimization, learning and inference particularly using them to understand, model and combat noise and uncertainty in real world applications. His current research centers on the application of novel, theoretically well motivated optimization algorithms to allocate natural resources with a focus on problems that arise in next generation smarter energy management systems.
His thesis research at Carnegie Mellon was in the application of information and coding theory, detection, probability theory and inference algorithms to a variety of sensing systems such as sensor networks, mobile robots, biological screening and drug discovery. He used information theoretic techniques to study the fundamental limits of inference in measurement systems. He also studied these problems from an algorithmic point of view and developed practical algorithms, aided by the representation of the system as a graphical model. One striking algorithmic insight that arose in this work is the computational cut-off phenomenon in sparse measurement systems. In contrast to the conventional accuracy-complexity trade-off, he showed that when a sufficiently large number of measurements are available, it is possible to have have a sharp theoretical and empirical performance transition, with algorithms that become both computationally efficient and accurate.
During his graduate studies he was fortunate to have the opportunity to collaborate with a number of bright people on problems ranging from target tracking, iris recognition, speaker recognition, multi-source separation and codes for next generation memory systems. He helped develop fast inference algorithms for target tracking combining methods from graphical models and computational signal processing. He worked on a conditional random field model of distortions in images and built a very accurate iris recognition system. He demonstrated techniques to build models and separate speakers in noisy recordings of phone and meeting conversations without any prior data on the speakers. Finally, he showed how to use simulation based optimization algorithms to design efficient codes, that used a novel iterative encoding scheme, particularly suited to future memory systems.
He is a proud recipient of the National Talent Search (NTSE) and the Jawaharlal Nehru (JNCASR) scholarships from the the government of India during his undergraduate studies.