Dr. Jayadeva is a Professor in the Department of Electrical Engineering at IIT Delhi. He currently holds the Microsoft Chair at IIT Delhi.

He has been a speaker on the IEEE Computer Society Distinguished Visitor Programme, and is a recipient of awards from the Indian National Academy of Engineering and the Indian National Science Academy. He was a URSI Young Scientist at the General Assembly in Lille, France (1996), and received the Sir J.C. Bose Young Scientist title from the Indian Council of the URSI. He visited the the Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology in 1997 as a BOYSCAST fellow. He spent a sabbatical year as a visiting Researcher at IBM India Research Laboratory, from July 2006. One of his papers in Neurocomputing was listed on the Top25 hotlist; he is a recipient of best paper awards from the IETE Journal of Research, and two other conference papers. He holds a US Patent on A/D conversion, another on assessing pronunciation abilities, and is the co-author of the book “Numerical Optimization and Applications”. He has served on the Steering and Program Committees of several international conferences. He was the tutorial chair for PrEMI 2009, delivered an invited talk at ICICS, Singapore, 2011, and was a keynote speaker at the IEEE SSCI 2013, Singapore.

His research interests include Machine Learning, Optimization, Swarm Intelligence, and the implementation of these in VLSI. Amongst recent notable work by him and his collaborators and students, is the Twin Support Vector Machine, which is cited by  publications and is the subject of an overview article in AI Review. Other significant work includes a new learning algorithm for SVMs, called 1SMO, that is twice as fast as the conventional SMO; SVMs that learn functions and their partial derivatives, fuzzy SVMs for time series forecasting, and neural networks for predicting textile properties.

New work on finding learning machines with minimal VC dimension shows that state-of-the-art solutions may be far from optimal, and that a new approach, called the Minimal Complexity Machine, can learn more efficient solutions with less computational effort.

Other interesting work includes a new global optimization algorithm, GOSAM, that combines SVM learning and local search. GOSAM uses knowledge from previous local minima to predict better starting points, in an attempt to find the global optimum; it outperforms other reported approaches such as Particle Swarm Optimization, Genetic Algorithms, and Simulated Annealing, in terms of time (by 2-3 orders) and solution quality.

Jayadeva and his students were amongst the first worldwide to realize a SVM application on a VLSI chip; a SVM classifier based A/D converter was fabricated in 180nm CMOS in 2006. The design also demonstrates a radically new design approach for self-calibrating analog circuits.

Jayadeva and his students showed that nonlinear analog circuits can perform tasks such as sorting in ideal sub-linear (1/N) time, and that practical circuits can be built for sorting in constant time. He also showed how networks of modified Phase Locked Loops can be used to solve optimization problems like graph colouring.

Very recent work includes the first theoretical results in the literature on convergence problems with traditional ant colony optimization methods. Employing Polya urn processes, this work proved that conventional ant algorithms cannot overcome initial bias beyond a point. A new algorithm called EigenAnt has been devised, that provably converges to the shortest path regardless of initial conditions, noise, and dynamic topology changes. Jayadeva presented this work as a keynote speaker at the 2013 IEEE Symposium Series on Computational Intelligence in Singapore. A new chip taped out in February 2013 includes an EigenAnt implementation, and to the best of knowledge, is the first ant colony optimization based on-chip router to be realized on silicon.

To the best of our knowledge, EigenAnt is the only ant colony optimization algorithm that provably converges to the shortest path.

Two conference and one journal paper of Jayadeva have received best paper awards, and additionally, one journal paper was listed on the Top 25 Hotlist of the journal Neurocomputing. He is co-author of a book on Numerical Optimization with Applications.

He has been issued two US patents – one on A/D conversion, and another on a method for assessing pronounciation abilities.