Meghana Kshirsagar


Graduate student

Language Technologies Institute

Gates Hillman Center (GHC) 5513


I am currently a post-doc at MSKCC in Christina Leslie's group, prior to which I did a brief stint at IBM Research, in the Machine Learning group (headed by Naoki Abe) working with Aurelie Lozano and Eunho Yang. I graduated from LTI in Aug 2015, where I was advised by Prof. Jaime Carbonell and Prof. Judith Klein-Seetharaman. I am interested in the broad area of machine learning for computational biology. My thesis focussed on multitask and transfer learning methods for host-pathogen protein-protein interaction prediction.


Before computational biology I have had not-so-brief stints in the areas of Information Retrieval for structured data (during my Masters at IIT Bombay) and Information Extraction from unstructured data (at Yahoo! Labs). I worked with S. Sudarshan at IIT Bombay and with Rajeev Rastogi, Srinivasan H. Sengamedu and Shailesh Kumar at Yahoo! Labs. I have also interned at IBM research labs, Delhi with Prasan Roy and Venkatesan Chakravarthy.

In summer of 2012, I interned at Microsoft Research, Redmond and I worked on a problem in the broad area of Genetics with Hoifung Poon. I was a TA for Machine Learning in the Spring of 2012, with Roni Rosenfeld. In the Spring of 2011, I TA-ed Data Mining and Information Retrieval with Jaime Carbonell.

Reviewing

Program Committee: BMC Genomics, IJCAI 2016, WWW 2017 Posters, Workshop for ML in Comp Bio 2016

Sub-reviewer: Machine Learning Journal 2015, AAAI 2014, Bioinformatics, BMC Genomics, WWW 2010, SIGMOD 2010

Publications

Proceedings and Journals


"Multitask matrix completion for learning protein interactions across diseases", Meghana Kshirsagar, Jaime Carbonell, Judith Klein-Seetharaman and Keerthiram Murugesan. RECOMB 2016 [Paper] [Code] (awaiting publication in JCB)


"Frame-Semantic Role Labeling with Heterogeneous Annotations", Meghana Kshirsagar, Sam Thomson, Nathan Schneider, Jaime Carbonell, Noah A. Smith and Chris Dyer, Assoc. for Computational Linguistics (ACL) 2015 [Paper (short paper)] [Code] [Slides] (part of the SEMAFOR system)


"Techniques for transferring host-pathogen protein interactions knowledge to new tasks", Meghana Kshirsagar, Sylvia Schleker, Jaime Carbonell and Judith Klein-Seetharaman, Frontiers in Microbiology, 2015. [Paper] [Code]


"Comparing human–Salmonella with plant–Salmonella protein–protein interaction predictions", Sylvia Schleker, Meghana Kshirsagar and Judith Klein-Seetharaman, Frontiers in Microbiology, 2015. [Paper]


"Multi-task learning for Host-Pathogen protein interactions", Meghana Kshirsagar, Jaime Carbonell and Judith Klein-Seetharaman, Bioinformatics 2013 [Paper]
[Supplementary, Data, Code]


"Techniques to cope with missing data in host-pathogen protein interaction prediction", Meghana Kshirsagar, Jaime Carbonell and Judith Klein-Seetharaman, Bioinformatics 2012 (also presented at ECCB 2012, Basel). [Paper] [Supplementary Info]


"Virus interactions with human signal transduction pathways", Zhongming Zhao, Junfeng Xia, Oznur Tastan, Irtisha Singh, Meghana Kshirsagar, Jaime Carbonell, Judith Klein-Seetharaman, I. J. Computational Biology and Drug Design 4(1): 83-105 (2011). [Paper]


"High Precision Web Extraction using Site Knowledge", Meghana Kshirsagar, Rajeev Rastogi, Sandeepkumar Satpal, Srinivasan Sangamedu and Venu Satuluri, Conference On Management of Data (COMAD) 2010 (Best Paper Award) [Paper]


"Keyword search on external memory data graphs", Bhavana Dalvi, Meghana Kshirsagar and S. Sudarshan, Very Large Databases (VLDB) 2008. [Paper]


Workshop papers, Posters


"Iteratively Regrouped Lasso: learning group structure in genome wide association studies in crops", Meghana Kshirsagar, Aurelie Lozano and Eunho Yang, Workshop on Data Science for Food, Energy and Water, Knowledge Discovery and Data Mining (KDD) 2016


"Automated Sorghum Phenotyping and Trait Development Platform", Mitch Tuiinstra, Cliff Weil, Addie Thompson, Chris Boomsma, Melba Crawford, Ayman Habib, Edward Delp, Keith Cherkauer, Larry Biehl, Naoki Abe, Meghana Kshirsagar, Aurelie Lozano, Karthikeyan Natesan, Peder Olsen, Eunho Yang, Workshop on Data Science for Food, Energy and Water, Knowledge Discovery and Data Mining (KDD) 2016


"Leveraging Heterogeneous Data Sources for Relational Semantic Parsing", Meghana Kshirsagar, Nathan Schneider and Chris Dyer, Assoc. for Computational Linguistics (ACL) 2014 workshop on Semantic Parsing


"Multisource transfer learning for host-pathogen protein interaction prediction in unlabeled tasks", Meghana Kshirsagar, Jaime Carbonell and Judith Klein-Seetharaman, NIPS Workshop on Machine Learning for Computational Biology, 2013 [Paper] [Talk Slides]


"Confident prediction of Salmonella-human protein-protein interactions" Schleker S., Nouretdinov I., Garcia-Garcia J., Oliva B., Meghana Kshirsagar, Klein-Seetharaman J. and Gammerman A, ECCB 2012.


"Transfer learning based methods towards the discovery of host-pathogen protein-protein interactions", Meghana Kshirsagar, Jaime Carbonell and Judith Klein-Seetharaman, ISMB 2012. [Poster]

Personal

If you have an old but working laptop/PC to donate to a needy child, please contact me! Here's the website of our campaign: CMU Laptop Rehab

Some of my hobbies are reading science-fiction, birding, photography, kayaking, playing board-games. I have trained in Kathak, an Indian classical dance form. I also love to ice-skate and ski (still a novice) in Pittsburgh winters. These days I'm learning the piano.