Inaugural Project: Modeling the Emergence of Proto-Cognition

The first research project developed by YHouse proposes to develop a complete bottom-up model of the origins of cognition starting from a chemical substrate.


The emergence of cognition marks a great evolutionary transition in the behavior of living systems. Whereas before, organisms’ survival depended on their genetic pre-dispositions, cognition allowed them to alter their decision making process as a consequence of their learning and development. We do not yet know how this transition to cognition occurred, and what underlying principles were responsible for it. The purpose of this inaugural project is to develop computational models of biological individuals that engage with other biological individuals and with their environments in a minimally cognitive way. Solving the theoretical challenges of developing this model will deepen our understanding of cognition and how it emerged, the lessons of which could apply broadly in the fields of cognitive science and synthetic cognition.


The goal of this project is to identify conditions that led protocells to assimilate relevant information from their environment to improve their own survival. Protocells are minimal chemical systems that have been extensively studied in context of the origin of life. By extending simulations of protocells to include minimally cognitive behaviors, both individually and in social contexts, we can develop the necessary concepts and methodologies for studying the emergence of cognition.

“What I cannot create, I do not understand,” said Richard Feynman. Computational models have the power to simplify the physics of the real world, while preserving essential functional properties. By building computational models of cognitive systems, we can come to understand their principles, and develop a theory of how cognition might emerge. The resulting cognitive systems can be systematically analyzed — simulated thousands of times in varying experimental conditions, and with analytic tools probing the state of every relevant component.

This project will develop in several stages. First, we will develop models of protocells that self-assemble in an artificial chemistry. We will quantify the information flow between the protocells and their environments, and study the structure of these interactions. Next, we will study the interaction among multiple protocells put in a same environment. We will quantify the information flows between these individuals, and the integrated behavior of the resulting social system. Finally, we will identify conditions in which protocells construct a network of social interactions that help them learn adaptive behavior throughout their lifetime.

By demonstrating the emergence of minimal forms of cognition in a computational model, we plan to provide a foundation for a theory on the emergence of cognition. This can later be generalized to more complex classes of cognition, allowing us to build cognitive robots or extend the cognition of real living systems.


Olaf Witkowski’s background is in building large-scale simulations and modeling the origins of communication in complex systems. He has developed analysis techniques based on information theory, machine learning and artificial neural networks.

Eran Agmon is an expert in modeling the emergence and stability of protocells from chemistry. His research uses tools from network theory and dynamical systems theory to analyze the processes of adaptation and deterioration.

Piet Hut has decades of experience with large-scale simulations in astrophysics, specifically of galaxies and star clusters, and has developed several algorithms that have been used extensively in astrophysics as well as in other areas such as protein folding.