Steffen will be visiting From Jan 21 to Feb 1 as part of the ADSI Visiting Faculty Program. He will be sitting in the ADSI faculty office, room CSE2 315. During his visit, Steffen will be giving a ML lunch seminar on Thurs Jan 23rd at noon. The details are follows:
Title: Oblivious data for kernel methods
Abstract: I’ll present an approach to reduce the influence of sensitive features in data in the context of kernel methods. The resulting method uses Hilbert space valued conditional expectations to create new features that are close approximations of the original (non-sensitive) features while having a reduced dependence on the sensitive features. I’ll provide optimality statements about these new features and a bound on the alpha-mixing coefficient between the sensitive features and these new features. In practice, standard techniques to estimate conditional expectations can be used to generate these features. I’ll discuss a plug-in approach for estimating conditional expectation which uses properties of the empirical process to control estimation errors.
Short bio: Steffen is an assistant professor in the department of mathematics and statistics at Lancaster University, UK. He joined the department in autumn 2014. His main area of research is large-scale machine learning with a focus on the interplay of kernel methods, convex optimization in Hilbert spaces and empirical process theory. Prior to joining Lancaster University, Steffen was a postdoctoral researcher in computer science at University College London. He completed his PhD in machine learning at the Technical University Berlin.