Speaker:
Vahab Mirrokni 

Session 3

Time: 9.00 AM, Feb 1st, CSL B02

Abstract:

As an important component of any ad serving system, online capacity (or budget) planning is a central problem in online ad allocation. I will survey primal-based and dual-based techniques borrowed from online stochastic matching literature and report theoretical approximation guarantees and practical evaluations of these algorithms on real data sets. Finally, inspired by practical applications, I will discuss a two new results in the area: (i) simultaneous approximation for both adversarial and stochastic models and recent theoretical results in this context, and (ii) computationally efficient mechanisms for dealing with budget constraints in the presence of strategic agents.