Decisions & Reinforcement Learning

Chibany learns to act. Believing is not enough: autumn brings the health kick, and every belief now has to cash out as a choice. This Part goes from beliefs to actions — losses and decisions, planning across time (MDPs), learning to act without a map (Q-learning), and then the great reversal: watching someone else act and inferring their goal (inverse RL, theory of mind), acting when the world itself is hidden (POMDPs and the Tiger problem), and the modern frontier where these ideas train today’s largest models.

graph LR
    A[Decision<br>Theory] --> B[MDPs]
    B --> C[Q-Learning]
    C --> D[Inverse RL 🧠]
    D --> E[POMDPs<br>& Belief]
    E --> F[Modern RL &<br>World Models]

Chapters

• Statistical Decision Theory: From Beliefs to Actions
• Markov Decision Processes: Planning When You Know the World
• Q-Learning: Acting Without a Map
• Inverse RL: Reading Goals from Behavior
• POMDPs and Belief: Inferring a Hidden World
• Modern RL: Preferences, World Models, and Machine Minds

This project is generously funded by the Japanese Probabilistic Computing Consortium Association (JPCCA).