Title: Simple models for optimizing driver earnings in ride-sharing platformsAnalysis for Machine Learning

Speaker: Evimaria Terzi

Abstract:On-demand ride-hailing platforms like Uber and Lyft are helping reshape urban transportation, by enabling car owners to become drivers for hire with minimal overhead. Such platforms are a multi-sided market and offer a rich space for studies with socio-economic implications.

In this talk I am going to address two questions:

In the absence of coordination, what is the best course of action for a self-interested driver that wants to optimize his earnings?
In the presence of coordination, is it possible to maximize social welfare objectives in an environment where the objectives of the participants (drivers, customers and the platform) are (often) misaligned?

We will discuss the computational problems behind these problems and describe simple algorithmic solutions that work extremely well in practice. We will demonstrate the practical strength of our approaches with well-designed experiments on novel datasets we collected from such platforms.

Bio: Evimaria Terzi is an Associate Professor at the Computer Science Department of Boston University. She joined the department in 2009 after spending two years as a Research Staff Member at the IBM Almaden Research Center. Evimaria got her PhD from the University of Helsinki in 2007. Her research interests span a big range of data-centered problems. Currently she is interested in problems related to team formation, ride-sharing platforms as well as recommender systems. He research is funded by multiple NSF grants and she was a recipient of the NSF CAREER award (2013) and Microsoft Faculty Fellowship (2010).