| I am currently a first year graduate student at the Department of Electrical Engineering and Computer Science, UC Berkeley. The easiest way to contact me is through email at or . |
Interests
I am broadly interested in Machine Learning and Artificial Intelligence. More specifically, my current research interests include:
- Statistical Learning Theory - I'm pretty interested in studying the generalization ability of algorithms. Recently I've also started working on Reductions between learning problems and their relationships in general. I am also working on learning of kernels from data.
- Information Retreival and Mining - I've done work in Web Mining and Text Classification. Currently I'm doing work in learning ranking for web pages using user preferences in setups that generalize both classical Graph based approaches and Ordinal Regression.
If you want to read more, take a look at my resume. (You need to manage with the pdf now, too lazy to put up an HTML :)
NOTE: The resume and the contents that follow haven't been updated since I left IIT. I'll hopefully get down to it sometime.
|
Publications
- Information-theoretic lower bounds on the oracle complexity of convex optimization
with Peter Bartlett, Pradeep Ravikumar and Martin Wainwright
To appear in NIPS 2009.
- A Stochastic View of Optimal Regret through Minimax Duality
with Jake Abernethy, Alexander Rakhlin and Peter Bartlett
arXiv preprint, to appear in COLT 2009.
- Matrix Regularization techniques for online multitask learning
with Alexander Rakhlin and Peter Bartlett
EECS Department, University of California, Berkeley, Tech. Rep. UCB/EECS-2008-138, Oct. 2008.
- Message-passing for graph structured linear programs: Proximal projections, convergence and rounding schemes
with Pradeep Ravikumar and Martin Wainwright
In ICML 2008. Longer version as Department of Statistics, University of California, Berkeley, Technical Report 765.
- An Analysis of Inference with the Universum
with Fabian Sinze, Olivier Chapelle and Bernhard Schölkopf
In NIPS 2007
- Learning Random Walks to Rank Nodes in Graphs
with Soumen Chakrabarti
In ICML 2007
- Learning Parameters in Entity-relationship Graphs from Ranking Preferences
with Soumen Chakrabarti
In ECML/PKDD 2006
- Learning to Rank Networked Entities
with Soumen Chakrabarti and Sunny Aggarwal
In SIGKDD 2006
- Sentiment Analysis: A New Approach for Effective Use of Linguistic
Knowledge and Exploiting Similarities in a Set of Documents to be
Classified.
with Pushpak Bhattacharyya
International Conference on Natural Language Processing(ICON),
IIT Kanpur, India, December 2005
- Augmenting WordNet with Polarity Information on Adjectives
with Pushpak Bhattacharyya
3rd International Wordnet Conference(GWC 06), Jeju Island, Korea, South Jeju (Seogwipo).
|
Courses Done
- Statistical Foundations of Machine Learning
- Information Retrieval and Mining
- Data Mining and Data Warehousing
- Probability Theory and Statistics
- A First Course in Optimization
- Functional Analysis
- Artificial Intelligence
- Design and Analysis of Algorithms
- Data Structures and Algorithms
- Theory of Computation
- Discrete Structures
- Logic Design
- Database and Information Systems
- Formal Methods in Computer Science
- Software Systems Laboratory
|
Work
Here I will be describing of some of the important projects, talks, seminars etc. done by me.
Ongoing Projects
Seminars and Talks
- Designing Kernels on Structured and Unstructured Input Spaces - This was a seminar that I completed in the sixth semester under the guidance of Prof. Soumen Chakrabarti. More details on this can be found here.
Projects Completed
Attribute Value Extraction from Semi-Structured Data(Spring 2006) - The aim of the project was to mine attribute value information from semi-structured data like that on e-commerce websites. We studied a DCFG based appraoch to this problem.
Encyclopaedic Article Relational Database(Autumn 2005) - We made an open source encyclopaedia similar to Wikipedia. Variety of advanced features like an option to specify user customized queries for every document, easy to use search interface, user account based database administration and content monitoring etc. were provided. This was implemented using Oracle SQLPlus using JDBC and JSP for user interfaces.
Sentiment Analysis of Movie Reviews(Summer 2005) - This project was done as a part of the UnderGraduate Research Oriented Project(UROP) under the guidance of Prof. Pushpak Bhattacharyya. A combination of supervised and unsupervised techniques was used to classify movie reviews as good or bad. Wordnet based techniques were used to identify the strength of adjectives in a good vs bad classification. High classification accuracy was obtained.
Information Retreival from Homepages and Web Document Classification(Spring 2005) - We designed heuristic approaches to mine information like research interests, hobbies, department etc. from university homepages. We used NLP based techniques for handling context shifts and polarity analysis. A second module used SVM with bag-of-words features to classify local repository of web pages collected from Open Directory Project(ODP). High classification accuracy was demonstrated.
Keyword search in a set of Hyperlinked Documents(Autumn 2004) - A scalable design for a search engine was implemented in C++. The crawling was done on a local repository of hyperlinked files. Result ranking was done using a combination of PageRank and text match with the query.
|
Links
Some useful and some fun links.
|
