Homepage-Rajhans Samdani


Rajhans Samdani Graduate Student Department of Computer Science University of Illinois at Urbana-Champaign E-mail: Combine @illinois.edu and rsamdan2 appropriately. Other e-mail: Same for @gmail.com and my-first-name.

About me

I am a fourth year PhD student and a sixth year graduate student (add 2 for MS) working with Prof. Dan Roth. I work on Machine Learning techniques for Natural Language Processing. More specifically, I work on supervised and un(semi-)supervised learning for structured output prediction for NLP applications. I am also interested in domain adapation and transfer learning for text data.

I have done research internships at Yahoo! Labs, Microsoft Research, and Google Research in the summers of 2008, 2010, and 2012, respectively.

I received M.S. in Computer Science from the same Department in May, 2009. Before that, I obtained a B. Tech degree from the Department of Computer Science at IIT Bombay.

Resume

I am looking for jobs. Here is my CV (updated Dec 18, 2012.)

Publications

A Framework for Tuning Posterior Entropy in Unsupervised Learning with Ming-Wei Chang and Dan Roth. ICML workshop on Inferning: Interactions between Inference and Learning, (2012).

Efficient Decomposed Learning for Structured Prediction with Dan Roth. ICML 2012.

Supplementary Material Containing Proofs.

Unified Expectation Maximization with Ming-Wei Chang and Dan Roth. To appear in NAACL 2012.

Illinois-Coref: The UI System in the CoNLL-2012 Shared Task with Kai-Wei Chang, Alla Rozovskaya, Mark Sammons, and Dan Roth. CoNLL Shared Task 2012.

Domain Adaptation with Ensemble of Feature Groups with Wen-tau Yih. IJCAI 2011.

Constrained Conditional Models For Information Fusion with Gourab Kundu and Dan Roth. Fusion 2011.

Inference Protocols for Coreference Resolution with Kai-Wei Chang, Alla Rozovskaya, Nick Rizzolo, Mark Sammons, and Dan Roth. CoNLL Shared Task 2011.

Automatic Model Adaptation for Complex Structured Domains with Geoffrey Levine, Gerald DeJong, Li-Lun Wang, Shankar Vembu, and Dan Roth. ECML PKDD 2010.

Learning Multi-linear Representations of Distributions for Efficient Inference with Dan Roth. 2009. ECML-PKDD 2009 and Machine Learning Journal. Amongst 14 out of 600 papers to be published in a special Machine Learning journal edition of ECML PKDD 2009.