Optimal Transport (OT) is a classic area in probability and statistics for transferring mass from one probability distribution to another. Recently OT has been very successfully used for domain adaptation in many applications in computer vision, texture analysis, tomographic reconstruction and clustering. We introduce a new regularizer OT which is tailored to better preserve the class structure. We give the first theoretical guarantees for an OT scheme that respects class structure. We give an accelerated proximal--projection scheme for this formulation with the proximal operator in closed form to give a highly scalable algorithm for computing optimal transport plans. Our experiments show that the new regularizer preserves class structure better and is more robust compared to previous regularizers.
18/04/2019
12am, room G50, building G, 3rd floor (viale Regina Elena 295/B).
Devdatt Dubhashi is Professor in the Department of Computer Science and Engineering at Chalmers University, leading the algorithms and machine learning group. He is also Chief Scientist at Machine Intelligence Sweden AB, a startup that bridges from research to innovation in AI powered technologies. He received his B.Tech. in Computer Science and Engineering from the Indian Institute of Technology (IIT) Delhi and his masters and Ph.D. in Computer Science from Cornell University USA. He has held positions at the Max Planck Institute for Computer Science Saarbrueken Germany and at BRICS (Basic Research in Computer Science) a center of the Danish National Science Foundation at the University of Aarhus Denmark. He led the project "Data Driven Secure Business Intelligence" funded by SSF and was an expert consultant for the OECD on Data Driven Innovation. He was a Vinnova and Marie-Curie Fellow He was an invited speaker in 2017 at the SICS Data Science Day, Vehicle Electronics and Connected Services and at the European Meeting of Statisticians.