1 December 2020 Tue 4 pm

< Previous

Next >

While complex systems are manifested in different forms (e.g. human brains, ecosystems, the Universe, quantum dots, nanorobots, etc), they can resemble one another in the way in which information is handled and processed. Information theory thus provides a powerful conceptual framework and method to analyze and understand seemingly different complex systems. In this talk, I will introduce and discuss a new information geometrical method that we have developed recently. The rationale for this method is that complex systems are time-varying and that it is thus useful to use the notion of distance to describe this time-varying dynamics. Technically, this distance (information length) is a path-dependent quantity and measures the number of statistically different states that a system evolves through along the path. I will discuss the merit of this method in understanding attractor structure, force-geometry link, chaos, correlation, etc.