A Thermodynamic Model for the Emergence of Intelligence
The most beloved elements of our societies are often products of our intelligence, thus our interest in reproducing it. The advent of natural intelligence as an emergent phenomena has recently received worthy attention. A fundamental understanding of this emergence is required for the successful advancement of artificial intelligence (AI) technology. Recent studies suggest that intelligent behaviours emerge in accordance with the second law of thermodynamics, which states that the entropy of a system always increases. These studies conclude that an intelligent body will make decisions that maximize both current and future opportunities for action.
We aim to investigate this claim using a randomized Markov chain model applied to a state vector. Specifically, using R we will multiply a state vector p(t) by a transition matrix A at each time step t. The state vector and transition matrix serve as proxies for an intelligent body and the entropy-maximizing force that acts on the body on a large scale respectively. We will calculate entropy (S) at each time step using the following formula:
This will allow us to record changes in entropy over time, determining how closely the system follows the second law. If it is in fact true that intelligence emerges from the second law, we expect to see an increase in entropy over time. This research will contribute to current literature regarding intelligence as an emergent phenomena that can be characterized by thermodynamic principles. Furthering our understanding of the emergence of natural intelligence can accelerate the advancement of AI.