Marco Antonio Pedroso
MIT EECS - eBay Inc. Undergraduate Research and Innovation Scholar
Analysis of Auctions and Pricing with little information about buyers valuations
How much should you charge for a subscription to Netflix or the New York Times? On this project we study classical and novel algorithms for setting prices for goods, given that we have much uncertainty about how much users are willing to pay. This uncertainty is often modeled as a probability distribution from which users values for the goods are drawn. But even when this distribution exists, the seller of the good may not know it or may only know partial information about it. Different theoretical algorithms have been proposed for the problem of maximizing revenue with uncertain demand, and we wish to understand how they compare in practice. My role will be to help implement these algorithms and compare their performance.
I have worked with Professor Medard on network coding algorithms for data storage in content distribution networks to improve the energy efficiency of the process. Working in the R&D of a broadcasting company, I optimized the parameters for encoding videos using self-learning algorithms, analyzing the quality and speed trade-off.