Docenturföreläsning med Denis Kleyko, datavetenskap

Title: Randomness and Computation

Abstract:
Traditionally, in computer science, it is not so common to think about randomness as a phenomenon facilitating computation (cryptography being an exception) but rather as a hurdle. There are, however, schools of thought that have been developing approaches to harvest randomness for computation. A crucial idea is that randomness needs to be combined with some structure-imposing mechanisms as otherwise there is no basis for meaningful computations.

This talk will begin by providing several seminal examples of computing with randomness from the past. These cases will be used to introduce the idea of computing with high-dimensional random vectors – an area of presenter's studies. Throughout the talk we will examine several nontrivial connections that this area features to different corners of computer science such as sketching algorithms, kernel methods, neural networks, and neuromorphic computing. Finally, if time permits, we will look into the social aspects of building a community around research ideas by examining some of the recent consolidation efforts within the area.

Welcome!