Characterizing the Emergence of Agency
Information is a crucially important concept for our understanding of life. The emergence and evolution of life can be largely understood as a change in the way information has been stored, processed, and transmitted through time (Maynard Smith & Sathmary, 1995). This common theme runs from the origin of natural selection right up to the emergence of human society. Our goal is to understand, from a physical/mathematical perspective, how these large-scale changes in the flow of information occur, and in particular, to understand the role that information dynamics can play in the emergence of complex structures similar to life.
In biology, living processes have been studied by breaking down their functions to gain insight into the compositional subsystems, and checking the results against empirical data collected from nature. On the contrary, our contribution constitutes a first step toward understanding the origin of complex life from a completely bottom-up approach, starting with simple theory on basic building blocks emerging from a simulated substrate, and understanding the underlying principles by construction from an information theory perspective. This computational approach will allow us to formalize the pathways to complex life, eventually leading us to generalize on the probability of finding life elsewhere in the universe, while arming us with unexpected tools to search for it.
We combine large-scale computer simulations, information theory, and game theory to factor out information patterns in living processes. Our work consists in 3 different parts:
- large-scale simulation—initialized with a set of artificial chemical compounds in a certain spatial arrangement, such that it reaches successive transitions in individuality, autonomy, agency, and social self-organization
- detection—characterizing these transitions in terms of informational patterns (e.g. characterizing agents within the system)
- analysis—quantifying the information flows with respect to the way they sustain the viability of information, taking trajectories leading to exploitation of free energy for their own maintenance
1. Eors Szathmary and John Maynard Smith (1995). The Major Evolutionary Transitions. Nature, 374(6519): 227–232.
2. Olivier Rivoire and Stanislas Leibler (2011). The Value of Information for Populations in Varying Environments. Journal of Statistical Physics, 142(6):1124–1166.
3. David Deutsch and Chiara Marletto (2014). Theory of Information Rewrites the Laws of Physics. New Scientist, 222(2970): 30–31.
4. Wolfgang Banzhaf and Lidia Yamamoto (2015). Artificial Chemistries. MIT Press.
5. Aubert-Kato, N., Witkowski, O., Hoel, E. and Bredeche, N. (2016). Towards Detecting the Emergence of Agency in Evolved Artificial Chemistries. Artificial Life XV: Late- Breaking Proceedings of the Fifteenth International Conference on the Synthesis and Simulation of Living Systems, 20–21.