A segway into information theory

Information theory may be useful for understanding active inference. As a minimum it offers some alternative perspectives on the quantities used in the active inference theory such as surprise, KL divergence, and entropy. This post provides a very short introduction to information theory.

Active inference lecture notes

I have set out to gain some insights into the active inference theory that provides a unified framework for perception, learning, decision making, and action. I will share my “lecture notes” combined with my own comments in this and future posts. My focus is on the mathematical models and the necessary algorithms.

Posted In AI