Integrated Information Theory with PyPhi: Testing and Improvement Strategies

The study of consciousness has increased in relevance in recent years in the scientific community. In the same line, integrated information theory (IIT) is making inroads into the understanding of consciousness. However, the enormous computational costs required make it difficult to apply IIT to experimental data. In this work we intend to explore and design efficient computational algorithms applying techniques such as Divide and Conquer and parallel computation that can provide some kind of solution to the challenges proposed by IIT based on optimizations and approximations used by PyPhi toolbox to reduce the complexity of the calculations. This software package allows users to easily study cause-effect structures of discret dynamical systems of binary elements, for causal analysis, serves as an up-to-date reference implementation of the formalisms of integrated information theory, and has been used in our research on efficient algorithms.