Dr. Thomas Parr - Active Inference Book

High and Low Roads

Active inference offers two distinct approaches: the 'high road' and the 'low road.' Thomas Parr explains that the high road is conceptual, focusing on the philosophical underpinnings of active inference, while the low road is mechanistic, emphasizing mathematical frameworks like Bayesian mechanics 1. Both paths aim to understand how living systems minimize surprise and free energy to resist entropy. Tim Scarfe highlights that the high road provides a helicopter view of how organisms resist entropic forces, whereas the low road delves into the mathematical intricacies of active inference 2.

Active inference puts action into perception, whereby perception is treated as perceptual inference or hypothesis testing.

--- Tim Scarfe

Ultimately, both roads converge on the idea that living systems optimize their interaction with the world.

   

Principles of Surprise

Surprise plays a central role in the framework of active inference, acting as a proxy for free energy. Thomas Parr describes surprise as the measure of how well a model fits the data, with less surprise indicating a better fit 3. This concept is crucial for understanding how organisms maintain homeostasis by minimizing surprise. Interestingly, Parr notes that organisms sometimes seek out surprise to better understand their environment, challenging the notion that minimizing surprise is the sole objective 4.

The best way of doing that is to actually gather as much information as you can about the world around you.

--- Thomas Parr

This dynamic interaction with surprise allows organisms to adapt and thrive.

   

Understanding Free Will

Active inference intersects intriguingly with concepts of free will and entropy. Thomas Parr discusses how agents, as open systems, interact with their environments, influencing entropy rather than being constrained by it 5. This interaction raises questions about free will, as agents make choices that affect their surroundings. Tim Scarfe adds that resisting entropic forces is essential for existence, yet agents also seek entropy to resolve uncertainties about their world 6.

Resisting entropic forces is what defines them, which gets us into the second law of thermodynamics.

--- Tim Scarfe

This duality highlights the complex dynamics of agency and free will within the active inference framework.