Nikhil Garg

Contact Information

Research Interests

Machine Learning, Computational Linguistics, Information Retrieval, Data Mining

Education

Université de Genève, Switzerland
Doctoral Student (2009 - Present)
Ecole Polytechnique Fédérale de Lausanne, Switzerland (EPFL)
Master of Science in Computer Science (GPA: 5.7/6) (2007 - 09)
Indian Institute of Technology, Delhi, India (IIT-Delhi)
Bachelor of Technology in Computer Science and Engineering (GPA: 8.5/10) (2002 - 06)

Work Experience

Yahoo! Research & Development, Bangalore, India
Software Engineer in Image Search (Jun 06 - May 07)
- relevance ranking of search results based on metadata, click logs, etc.
- filtering adult content

Internships

Idiap Research Institute, Switzerland
Research Assistant under Dr. Daniel Gatica-Perez and Prof. Hervé Bourlard (Feb 09 - Aug 09)
Worked on co-occurrence models for content-based automatic image annotation & retrieval, taking measures to address noisy and limited tagging problems in web image repositories such as Flickr.
International Computer Science Institute, Berkeley, USA
Visiting student in Speech group (Aug 08 - Jan 09)
Worked on summarization of meeting transcripts.
Ecole Polytechnique Fédérale de Lausanne, Switzerland
Research Assistant under Prof. Boi Faltings in AI Lab (Summer 2007)
Worked on an extension of a distributed constraint optimization algorithm, DPOP, using cryptography to preserve privacy of the participants.
IBM India Software Labs, Pune, India
IBM Extreme Blue Internship (Summer 2005)
Worked on an efficient communication mechanism for IBM Workplace Client Technology.

Teaching Assistantship

Structures de Données (Data Structures), Spring 2011
Méthodes empiriques et langages de script (Empirical Methods & Scripting Languages), Fall 2010
Systèmes Informatiques (Operating Systems), Fall 2009

Publications

Nikhil Garg and James Henderson, Temporal Restricted Boltzmann Machines for Dependency Parsing, ACL 2011 (short papers), Portland, USA
Nikhil Garg and Daniel Gatica-Perez, Tagging and Retrieving Images with Co-Occurrence Models: from Corel to Flickr, Workshop on Large-Scale Multimedia Retrieval and Mining in conjunction with ACM Multimedia 2009, Beijing, China
Nikhil Garg, Benoit Favre, Korbinian Riedhammer, and Dilek Hakkani-Tur, ClusterRank: A graph based method for meeting summarization, Interspeech 2009, Brighton, UK
Nikhil Garg and Ingmar Weber, Personalized, interactive tag recommendation for Flickr, ACM International Conference on Recommender Systems, 2008, Lausanne, Switzerland
Nikhil Garg and Ingmar Weber, Personalized tag suggestion for Flickr, World Wide Web Conference, 2008, Beijing, China
Nikhil Garg and Bipin Suresh, Predicting and ranking search query results based on click logs, US Patent filed in December 2007

Personal Projects

PicGuess
- A visual quiz asking users to guess the location of a photo. Uses geo-tagged photos from Flickr.
- implemented using python, django, java, mysql, Flickr API, Android SDK.
- source code

Programming Skills

Python, C++, Matlab, SML, Linux shell scripting