kalle21·Aug 22Monte-Carlo and Temporal-DifferenceWhen an agent doesn’t know how its environment works — it doesn’t have a model of transition probabilities or reward functions — it has to…
Cher Lim·Aug 19From Prediction to Decision: The Mathematics of Reinforcement LearningHow rewards, value functions and Bellman equations teach AI what to do
Meric_Gb·Jul 14Q-Learning & The Bellman EquationI was wondering how the hell gamma affects how long term the agent thinks
KAKANAKOU Miguel·May 25Teaching a Neural Network to Play Breakout: A Full DQN Implementation in JAXAfter around 9 hours of training, my DQN agent was playing Breakout with an average score of 42.52 and a peak of 90. Not…
Yeremia N.·May 9Lessons Learned: DecisionAmong numerous lessons learned over the past three decades, one valuable lesson is related to decisions. This post discusses decisions and…
Hari Prasannaa·Jan 13From “Memoryless” Models to Intelligent Decisions: A Story-Driven Guide to Markov, Bellman, and…Initially, I used to think “Markov” meant something almost insulting: memoryless. As if the model was saying, “I don’t care what happened…
Rohan Roy·Dec 8, 2025Basics Of Reinforcement Learning — Bellman Equations, Model-Free Learning, and Monte Carlo MethodsIn the previous post, we defined the MDP — the map of the world our agent lives in. But defining the world isn’t enough; the agent needs…
DhanushKumar·Dec 1, 2025Bellman Equation in Reinforcement LearningReinforcement learning is all about learning to make the best decisions through trial and error.An RL agent interacts with the environment…
![[Reinforcement Learning] Chapter 5](https://miro.medium.com/v2/resize:fill:160:106/1*6f6c-3L0P2QBwA3w3t5sxg.png)
![[Reinforcement Learning] Chapter 5](https://miro.medium.com/v2/resize:fill:320:214/1*6f6c-3L0P2QBwA3w3t5sxg.png)