Program : Bachelors of Technology
Advisor : Dr. Balaraman Ravindan
Areas of Interest : Machine Learning, Data Mining, Social Network Analysis, Natural Language Processing
Research Topic : Improving Aggregate Diversity of Recommendation Systems
Research in recommendation systems has turned away in recent years from just accurate prediction of ratings. One of the objectives that has come to light is aggregate diversity, which is a measure of how equal are the number of times each item is recommended, across users in the system. This would be useful in situations such as when a DVD rental service wants all videos on their shelves to be recommended to someone, or when assigning papers to review to experts, ensuring that all papers get assigned and in a load-balanced manner. My project involves identification of a suitable metric to evaluate aggregate diversity and improve the performance of systems on this metric, either using re-ranking methods with provable bounds or by directly optimizing the parameters of the system for this task.
Personal Webpage: Aishwarya Padmakumar
Courses Completed: Machine Learning, Data Mining, Social Network Analysis, Natural Language Processing, Indexing and Searching in Large Datasets, Introduction to Game Theory, Concurrent Programming
Courses Ongoing: Reinforcement Learning, Distributed Algorithms