Recent Projects

Truthful Multi-Armed Bandit Mechanisms for Multi-Slot Sponsored Search Auctions (Summer 2009):

This project was started while I was interning with Prof. Y Narahari, at the CSA department, IISc. The report is currently undergoing revisions for a conference publication. The project deals with the problem of characterizing truthful mechanisms for the pay-per-click sponsored search auctions extensively used by search engines. This work is essentially an extension to the paper "Characterizing Truthful Multi-Armed Bandit Mechanisms" by Moshe Babaioff, Yogeshwer Sharma and Aleksandrs Slivkins. They solve the problem of characterizing truthful MAB mechanisms for sponsored search auctions of a single slot and provide a regret bound on such mechanisms. We have attempted to extend their work to the multiple slot case. We show that assuming unrestricted click rates and no prior information about them, any truthful mechanism must satisy certain very rigid restrictions which result in a very high regret. We also characterize truthful mechanisms under some relaxed settings by making assumptions on the click rates, and experimentally obtain some regret estimates.
Scanning Tunneling Microscope - An Overview (Summer 2008):

This project was undertaken just after my 12th grade under Prof. Arindam Ghosh at the Physics Department, IISc. The project report consisted of the number following parts including both theoretical and experimental components.
- An introduction to the scanning tunneling microscope including a technical discussion on the quantum mechanical phenomenon tunneling.
- A complete summary of the modern tip preparation techniques.
- Using the STM to obtain several atomic scale images of HOPG.
- Finally, I built my own setup to prepare platinum tips, which gave tips of up to a few micron diameter consistently.

Back to Home