B. Hamzi: Opening remarks for the Symposium on ML and AIT, 07/22, Alan Turing Institute, UK

The goal of this symposium is to explore the interface between Algorithmic Information Theory (AIT) and Machine Learning (ML).

Algorithmic Information Theory studies complexity measures on data structures and the goal of this symposium is to bring together researchers applying tools of AIT, and particularly Kolmogorov Complexity, to ML in order to identify the domain of applicability of certain ML methods and explain, for example, why certain methods in ML will a priori work so well for certain problems or a priori won't work for other problems. Other questions of interest are the applications of Solomonoff induction, algorithmic probability approaches to probability predictions, links between information, learning, and data compression, developing/using ML algorithms for (better) compression/prediction trying to approximate Kolmogorov Complexity etc.

Organizing committee: Boumediene Hamzi, Kamaludin Dingle, Marcus Hutter, and Robert MacKay.