Our brand new course combining information theory, algorithmic complexity, and dynamical systems
Life From a Computational Perspective
Turing dies in 1954, one year after the discovery of the double-helical structure of DNA by James Watson and Francis Crick, but before biology's subsequent revolution. Neither he nor von Neumann had any direct effect on molecular biology, but their work allows us to discipline our thougths about machines, both natural and artificial.
Sydney Brenner, Nature 2012
This course provides a conceptual introduction to the new and exciting field of Algorithmic Information Dynamics focusing on mathematical and computational aspects in the study of causality. To this end, the course first covers key aspects from graph theory and network science, information theory, dynamical systems and algorithmic complexity in a tour the force to finally tackle causation from a model-driven approach removed from traditional statistics and classical probability theory. The course will venture into ongoing research to show exciting new avenues to uncharted territory.
After a conceptual overview of the main motivation and some historical developments, we review some preliminary aspects needed to understand the most advanced topics. These include basic concepts of statistics and probability, notions of computability and algorithmic complexity and brief introductions to graph theory and dynamical systems. We then dig deeper into the core of the course, that of Algorithmic Information Dynamics which brings all these areas together in harmony to serve in the challenge of causality discovery, the most important topic in science. Central to the course and the field is the theory of algorithmic probability that establishes a formal bridge between computation, complexity and probability.
Finally, we move towards new measures and tools related to reprogramming artificial and biological systems, applications to biological evolution, evolutionary programming, phase space and space-time reconstruction, epigenetic landscapes and aspects relevant to data analytics and machine learning such as model generation, feature selection, dimensionality reduction and causal deconvolution. We will showcase the tools and framework in applications to systems biology, genetic networks and cognition by way of behavioural sequences. Because of the wide scope of application students will be able apply the tools to their own data and own problems as we will be explaining how to do it in detail, and we will be providing all the tools and code for it.
Throughout the course, students will be given assignments that will go from the conceptual to the mathematical and computational intended to keep everybody engaged.
Instructors Hector Zenil and Narsis Kiani