ideasthesia

Sign in Subscribe

Ryan Collison

I write about cool science and share it here.

Agent at decision point with branching policy trajectories color-coded by expected free energy

Expected Free Energy: The Objective Function That Plans

Series: Active Inference Applied | Part: 3 of 10 Planning is not optional. Every system that persists across time must select actions that keep it viable. Bacteria swim toward nutrients. Predators stalk prey. You decide whether to read this sentence or scroll past it. These are not merely behavioral outputs—they

Active inference agent-environment loop visualization with generative model and prediction-action pathways

From Theory to Code: The Active Inference Implementation Revolution

Series: Active Inference Applied | Part: 1 of 10 In 2010, Karl Friston published a paper claiming that everything from bacterial chemotaxis to human consciousness operates on the same mathematical principle: organisms minimize the gap between what they expect and what they experience. By 2020, researchers were using these equations to

Active inference agent architecture as multi-scale coherence structure with hierarchical state space manifolds

Synthesis: Applied Active Inference and the Engineering of Coherence

Series: Active Inference Applied | Part: 10 of 10 Theory is beautiful. Implementation is revealing. This series began with a practical question: how do you actually build an active inference agent? We traced the mathematics of variational inference through generative models, expected free energy calculations, message passing architectures, and hierarchical planning.

Humanoid robot wireframe with luminous sensorimotor pathways and prediction-error flows

Robotics and Embodied Active Inference

Series: Active Inference Applied | Part: 7 of 10 In 1950, W. Grey Walter built a robot turtle named Elsie. She had two vacuum tubes for a brain, two motors, and a photocell that made her seek light. She also had a tendency to get stuck. When her battery ran low,

Software architecture visualization showing discrete and continuous computational pathways

PyMDP and RxInfer: The Active Inference Software Stack

Series: Active Inference Applied | Part: 5 of 10 There’s a gap between understanding the theory and actually building something that works. You can grasp the free energy principle, follow the math of expected free energy, even understand how message passing implements inference—and still have no idea where to

Factor graph visualization with variable and factor nodes connected by flowing message pulses

Message Passing and Belief Propagation in Active Inference

Series: Active Inference Applied | Part: 4 of 10 Building a generative model is one thing. Computing with it efficiently is another. Active inference agents need to infer hidden states, evaluate policies, and update beliefs in real time—all while minimizing something called variational free energy. The mathematics behind this is

Layered probabilistic world model visualization showing state-observation-action relationships

The Generative Model: Building World Models for Active Inference Agents

Series: Active Inference Applied | Part: 2 of 10 The difference between an active inference agent that works and one that fails catastrophically usually comes down to a single question: how well does its generative model capture the structure of its world? This isn’t philosophical abstraction. It’s the engineering