Kushal Dave

Ph.D. Student

Search and Information Extraction Lab

Language Technology research Center

IIIT - Hyderabad

E-mail: email

Phone : +91-9642538135



About Me

I am Ph.D. student working with Dr. Vasudeva Varma at International Institute of Information Technology, Hyderabad. I have received my bachelor's degree in Computer Science from GCOEA. Before pursuing my Ph.D. I worked for Persistent Systems Ltd. Nagpur as a Software engineer. After that, my desire to pursue higher studies brought me to IIIT-H where I joined as a Ph.D. student. I work in Search and Information Extraction Lab (SIEL) at the university.

My research focuses on retrieval and ranking of short textual documents for a given query. As a case study, I am looking the the short textual ads and studying the issues in retrieving the ads for the a given content (like, search query, web page content, other microblogging content). Specifically, I am investigating various semantic techniques that could help overcome the Vocabulary Mismatch in retrieving the short text for a given query. Currently, I'm investigating techniques ranging from Topic models to Conceptual Modeling of the text to retrieve the short text for the query. Conceptual modeling involves identifying the latent concepts behind the text and matching the query and the content in the concept space. I am also exploring the contribution of different other factors like role of the user (and/or advertiser, in the specific case of retrieving ads for a given content) in the retrieval and ranking of short text for a given content. As my research involves close interaction of user modeling, information retrieval and statistics, I work on various aspects of these areas. My research interests include Application of Machine Learning Techniques to problems in Information Retrieval and peripheral problems, Information Retrieval, Information Extraction etc.

This section of the homepage is updated more consistently at:

Recent News

Selected Publications

  • Kushal Dave, Vishal Vaingankar, Sumanth Kolar and Vasudeva Varma, Timespent Based Models for Predicting User Retention, WWW 2013 Rio-de-Janeiro Brazil
  • Ankit Patil, Kushal Dave and Vasudeva Varma, Leveraging Latent Concepts for Retrieving Relevant Ads For Short Text, ECIR 2013 Moscow Russia
  • Kushal Dave, Ankit Patil and Vasudeva Varma, Topic Models for Retrieving Relevant Ads, Workshop on Machine Learning in Advertising AdML 2012, KDD 2012, Beijing, China
  • Kushal Dave and Vasudeva Varma, Identifying Microblogs for Targeted Contextual Advertising, ICWSM 2012 AAAI Dublin, Ireland.
  • Kushal Dave Computational Advertising: Leveraging User Interaction & Contextual Factors for Improving Ad Relevance & Targeting, Doctoral Consortium, WSDM 2012, Seattle, U.S.
  • Kushal Dave and Vasudeva Varma, Towards Contextual Advertising in Micro-blogs, ICON 2011 Chennai, India
  • Kushal Dave, Rushi Bhatt and Vasudeva Varma, Modeling Action Cascades in Social Networks, ICWSM 2011 AAAI Barcelona Spain
  • Kushal Dave, Computational Advertising: Leveraging User Interaction & Contextual Factors for Improved Ad Retrieval & Ranking, PhD Symposium, WWW 2011, Hyderabad India
  • Kushal dave and Vasudeva Varma. Learning the Click-Through Rate for Rare/New Ads from Similar Ads, SIGIR 2010 Geneva, Switzerland.
  • Kushal dave and Vasudeva Varma. Pattern Based Keyword Extraction for Contextual Advertising, CIKM 2010 Toronto, Canada.
  • IIIT Hyderabad at TREC 2009, working notes of the Text Retrieval Conference, Million Query (1MQ) track.

Research

Research Interests:

  • Machine Learning
  • Information Retrieval
  • Information Extraction
  • Graph Theory

Current & Past Research Projects:

  • LSI modelling for Enterprise Search environment
  • TREC 2009 Million Query Track. Report
  • Parent-child pair extraction from Hindi corpus. Report
  • Search engine for Wikipedia text.

Links:

  • IREA / Yahoo! Summer School presentation on 'Learning CTR for New ads". ppt Detailed report
  • Lecture notes on Linear and Non-Linear SVM's. pdf
  • Notes on writing a 'good' research papers by Simon Jones MSR.

Other Stuff