Rishabh Iyer

My photo

Rishabh Iyer received his Ph.D at the University of Washington, Seattle, where he works with Prof. Jeff Bilmes. He did an internship in the summer of 2014 at the Machine Learning Group at Microsoft Research, Redmond, where he worked with Max Chickering and Chris Meek. In the summer of 2012, he did another internship at MSR, Redmond where he worked with Dr. Matthai Phillipose at the Mobility and Networking Research Group. Prior to coming to UW, he did his B.Tech at the Indian institute of Technology, Bombay, where he worked with Prof. Subhasis Chaudhuri. He also did an internship in 2010 at the Computer Science Department at Simon Fraser University, where he worked with Prof. Torsten Möller.

His research interests are at the intersection of Discrete optimization and Machine Learning, and particularly in subset selection problems in machine learning like summarization (including problems like summarizing image collections, document collections, news articles, videos etc.), data subset selection and active learning, semantic segmentation, feature selection, sensor placement, image correspondence etc. His work is focussed on a particular class of discrete optimization problems called Submodular Optimization, which on one hand are powerful, flexible and representable models in many real world applications, but on the other hand also admit algorithms with often strong theoretical guarantees. His work focusses on investigating certain theoretical characterizations of submodular functions and exploiting these in machine learning applications. One significant thread of his research revolves around providing faster, scalable and more practical algorithms for a large class of submodular optimization problems, and applying them to several real world problems. He is keenly interested in connecting theory with practice, and has successfully applied his algorithms to several real world applications in machine learning, computer vision, natural language processing, speech and information retrieval.

He recently received the Microsoft Research Fellowship award, in 2014. He also received the Facebook Fellowship award (declined in the favor of the Microsoft one), the Yang Outstanding Doctoral student award, and best paper awards at the top tier Machine Learning conferences, Neural Information Processing Systems (NIPS) and International Conference for Machine Learning (ICML). Here is a formal bio in third person.

For more information, please see his recent publications (DBLP, Google Scholar Profile), or you can contact him at rkiyer at u.washington.edu. Here is a link to his Curriculum Vitae.


Conference and Journal Publications

Ramakrishna Bairi, Rishabh Iyer, Ganesh Ramakrishnan and Jeff Bilmes, Summarizing Multi-Document Topic Hierarchies using Submodular Mixtures, To Appear In the Annual Meeting of the Association for Computational Linguistics (ACL), Beijing, China, July - 2015

Kai Wei, Rishabh Iyer, and Jeff Bilmes, Submodularity in Data Subset Selection and Active Learning, To Appear In Proc. International Conference on Machine Learning (ICML), Lille, France, June - 2015

Rishabh Iyer and Jeff A. Bilmes, Submodular Point Processes with Applications in Machine Learning In Proc. International Conference on Artificial Intelligence and Statistics (AISTATS), San-Diego, United States, May - 2015 PDF.

Yoshinobu Kawahara, Rishabh Iyer and Jeff Bilmes On Approximate Non-submodular Minimization via Tree-Structured Supermodularity, In Proc. International Conference on Artificial Intelligence and Statistics (AISTATS), San-Diego, United States, May - 2015 PDF.

Rishabh Iyer, Rushikesh Borse and Subhasis Chaudhuri, Embedding capacity estimation of reversible watermarking schemes, In Sadhana 39 (Part 6), 2014 PDF.

Sebastian Tschiatschek, Rishabh Iyer, Hoachen Wei and Jeff Bilmes, Learning Mixtures of Submodular Functions for Image Collection Summarization, In Advances of Neural Information Processing Systems (NIPS), Montreal, Canada, December - 2014 PDF.

Rishabh Iyer, Stefanie Jegelka and Jeff Bilmes, Monotone Closure of Relaxed Constraints in Submodular Optimization: Connections Between Minimization and Maximization, In Uncertainity in Artificial Intelligence (UAI), Quebec City, Canada, July - 2014 PDF, Extended Version.

Kai Wei, Rishabh Iyer, and Jeff Bilmes, Fast Multi-stage Submodular Maximization, In Proc. International Conference on Machine Learning (ICML), Beijing, China, June - 2014 PDF, Extended Version.

Rishabh Iyer and Jeff A. Bilmes, Submodular Optimization with Submodular Cover and Submodular Knapsack Constraints, In Advances of Neural Information Processing Systems (NIPS), Lake Tahoe, Nevada, December - 2013 PDF, Arxiv Extended Version, Slides, Video (Winner of the Outstanding paper award)

Rishabh Iyer, Stefanie Jegelka, and Jeff A. Bilmes, Curvature and Optimal Algorithms for Learning and Minimizing Submodular Functions, In Advances of Neural Information Processing Systems (NIPS), Lake Tahoe, Nevada, December - 2013 PDF, Arxiv Extended Version, Poster

Rishabh Iyer and Jeff A. Bilmes, The Lovasz-Bregman Divergence and connections to rank aggregation, clustering and web ranking , In Uncertainity in Artificial Intelligence (UAI), Bellevue, Washington, July - 2013 PDF, Extended Version, Slides (selected for oral presentation, 11% Acceptance rate).

Rishabh Iyer, Stefanie Jegelka, and Jeff A. Bilmes, Fast Semidifferential-based Submodular Function Optimization, In Proc. International Conference on Machine Learning (ICML), Atlanta, Georgia, June - 2013 PDF, Extended Version, Arxiv, Slides, Code, Video (Winner of the Best paper award).

Rishabh Iyer and Jeff A. Bilmes, Submodular-Bregman and the Lovasz-Bregman Divergences with Applications, In Advances of Neural Information Processing Systems (NIPS), Lake Tahoe, Nevada, December - 2012 PDF, Extended Version, Poster.

Rishabh Iyer and Jeff A. Bilmes, Algorithms for Approximate Minimization of the Difference between Submodular Functions, In Uncertainity in Artificial Intelligence (UAI), Catalina Islands, CA, August - 2012 PDF, Extended Version, Slides, Code (selected for oral presentation, 9% Acceptance rate).

Shah Ronak, Rishabh Iyer and Subhasis Chaudhuri, Object Mining for Large Video Data, In Proc. British Machine Vision Conference (BMVC), Dundee, Scotland, United Kingdom, August - 2011, PDF (selected for oral presentation - 8% Acceptance rate).

Workshops and Preprints

Rishabh Iyer and Jeff A. Bilmes, Near Optimal algorithms for constrained submodular programs with discounted cooperative costs, In NIPS Workshop on Discrete Optimization in Machine Learning (DISCML), Lake Tahoe, Nevada, December - 2014 PDF.

Rishabh Iyer and Jeff A. Bilmes, Submodular Point Processes, In NIPS Workshop on Discrete Optimization in Machine Learning (DISCML), Lake Tahoe, Nevada, December - 2014 PDF.

Yoshinobu Kawahara, Rishabh Iyer and Jeff Bilmes On Approximate Non-submodular Minimization via Tree-Structured Supermodularity, In NIPS Workshop on Discrete Optimization in Machine Learning (DISCML), Lake Tahoe, Nevada, December - 2014 PDF.

Rishabh Iyer, Stefanie Jegelka and Jeff A. Bilmes, Mirror Descent-Like Algorithms for Submodular Optimization, In NIPS Workshop on Discrete Optimization in Machine Learning (DISCML): Structure and Scalability, Lake Tahoe, Nevada, December - 2012 PDF, Poster.

Rishabh Iyer and Torsten Möller, A spatial domain optimization for sampling pointsets, MITACS Globalink Research Symposium, University of British Columbia, Canada - 2010. PDF

Professional Activities

Reviewer: International Conference of Machine Learning (ICML) 2013/2014, Neural Information Processing Systems (NIPS) 2013/2014, Uncertainty in Artificial Intelligence (UAI) 2015, Journal of Machine Learning Research (JMLR), Journal of Discrete Applied Mathematics (DAM).

Collaborators (Past and Present)

Jeff Bilmes, Stefanie Jegelka, Yoshinobu Kawahara, Kai Wei, Sebastian Tschiatschek, Chris Meek, Max Chickering, Patrice Simard, Ganesh Ramakrishnan, Matthai Phillipose, Bethany Herwaldt, Subhasis Chaudhuri, Torsten Möller.

Talks

Invited Talks: TOPS Seminar (UW), Yahoo! Machine Learning Lunch (UW), UW-MSR Machine Learning Day-2015, General Electric, IIT Bombay, IISc Bangalore, Microsoft Research Bangalore, IIT Gandhinagar (See related video)

Conference Talks: International Symposium on Mathematical Programming (ISMP-2015 -- Invited presentation), Neural Information Processing Systems (NIPS-2013), International Conference in Machine Learning (ICML-2013), Uncertainty in Artificial Intelligence (UAI-2012 and 2013), MITACS Globalink Research Symposium-2010.

Fun Stuff

My hobies include biking, hiking, squash, racquetball, tennis, cooking, meditation, and philosophical discussions. On weekends, I love to volunteer at different cultural organizations around Seattle. I am strong advocate of work-life balance! You can find me on facebook, LinkedIn and Google.