finding-theta technical blog: notes on search, inference, optimization, and reproducible engineering.https://www.findingtheta.com/enhttps://www.findingtheta.com/blog/reinforce-tactics-a-technical-compendium-and-analysis-of-large-language-model-performance-in-stochastic-strategy-environments/https://www.findingtheta.com/blog/reinforce-tactics-a-technical-compendium-and-analysis-of-large-language-model-performance-in-stochastic-strategy-environments/Explore a technical case study on 'Reinforce Tactics,' a new RL environment where traditional game theory outperforms the latest LLMs by a massive margin, exposing critical flaws in token-based reasoning.Thu, 01 Jan 2026 00:00:00 GMTprojectsreinforcement-learningreinforce-tacticsppolarge-language-modelshttps://www.findingtheta.com/blog/advanced-architectures-and-methodologies-in-visual-reinforcement-learning-a-technical-analysis-of-the-vizdoom-platform/https://www.findingtheta.com/blog/advanced-architectures-and-methodologies-in-visual-reinforcement-learning-a-technical-analysis-of-the-vizdoom-platform/Explores how the ViZDoom platform drives advancements in embodied AI, detailing the evolution from standard Deep Q-Networks to hierarchical architectures capable of mastering complex, 3D environmentsMon, 01 Dec 2025 00:00:00 GMTexperimentsreinforcement-learninghttps://www.findingtheta.com/blog/the-evolution-of-imagination-a-deep-dive-into-dreamerv3-and-its-conquest-of-minecraft/https://www.findingtheta.com/blog/the-evolution-of-imagination-a-deep-dive-into-dreamerv3-and-its-conquest-of-minecraft/Explore DreamerV3, the AI that taught itself to find diamonds in Minecraft, revolutionizing reinforcement learning with its powerful world model.Sat, 01 Nov 2025 00:00:00 GMTdeep-divesreinforcement-learningworld-modelshttps://www.findingtheta.com/blog/the-unseen-hand-guiding-a-virtual-drone-with-sparse-and-dense-rewards/https://www.findingtheta.com/blog/the-unseen-hand-guiding-a-virtual-drone-with-sparse-and-dense-rewards/Exploring training a drone with reinforcement learning, focusing on the trade-offs between sparse and dense rewards to achieve agile flight while avoiding unintended "reward hacking."Mon, 01 Sep 2025 00:00:00 GMTexperimentsmachine-learningreinforcement-learningroboticshttps://www.findingtheta.com/blog/ultimate-guide-to-contextual-bandits-from-theory-to-python-implementation/https://www.findingtheta.com/blog/ultimate-guide-to-contextual-bandits-from-theory-to-python-implementation/Discover the ultimate guide to contextual bandits, covering everything from core theory and key algorithms to a complete Python implementation with code for building powerful personalization and recommendation systemsFri, 01 Aug 2025 00:00:00 GMTdeep-divescontextual-banditshttps://www.findingtheta.com/blog/mastering-autonomy-a-comprehensive-guide-to-the-autodrive-ecosystem-and-reinforcement-learning/https://www.findingtheta.com/blog/mastering-autonomy-a-comprehensive-guide-to-the-autodrive-ecosystem-and-reinforcement-learning/A complete guide to the AutoDRIVE ecosystem. Learn to train self-driving car agents from scratch using reinforcement learning in this powerful Unity-based simulator.Tue, 01 Jul 2025 00:00:00 GMTexperimentsroboticshttps://www.findingtheta.com/blog/agentic-ai-the-autonomous-evolution-of-machine-learning-and-its-dawn-in-robotics/https://www.findingtheta.com/blog/agentic-ai-the-autonomous-evolution-of-machine-learning-and-its-dawn-in-robotics/Explore Agentic AI, the next frontier in machine learning. Discover how autonomous agents learn, act independently, and are revolutionizing the field of robotics.Sun, 01 Jun 2025 00:00:00 GMTdeep-divesroboticsagentshttps://www.findingtheta.com/blog/serving-up-some-robotics-setting-up-a-tennis-environment-in-mujoco/https://www.findingtheta.com/blog/serving-up-some-robotics-setting-up-a-tennis-environment-in-mujoco/Build a MuJoCo robot tennis simulation! Learn to set up a wall tennis environment, tackle physics/control challenges, understand its architecture, and improve it for robotics or reinforcement learning projects with MuJoCoThu, 01 May 2025 00:00:00 GMTprojectsmujococourtside-dynamicsreinforcement-learninghttps://www.findingtheta.com/blog/bridging-worlds-how-visual-language-action-models-are-teaching-robots-to-see-understand-and-act/https://www.findingtheta.com/blog/bridging-worlds-how-visual-language-action-models-are-teaching-robots-to-see-understand-and-act/Understand how Visual Language Action Models (VLAs) let robots follow commands by fusing AI vision and language for action, demonstrated with a conceptual Python robotics simulationTue, 01 Apr 2025 00:00:00 GMTdeep-divesroboticsvlahttps://www.findingtheta.com/blog/from-zero-to-dino-roar-teaching-a-t-rex-to-walk-with-mujoco-and-reinforcement-learning/https://www.findingtheta.com/blog/from-zero-to-dino-roar-teaching-a-t-rex-to-walk-with-mujoco-and-reinforcement-learning/A deep technical walkthrough of building a biomechanically faithful T-Rex in MuJoCo and training it to balance, walk, and hunt through three-stage curriculum learning with PPO — achieving a 96.7% bite success rate on commodity GPU hardware. Covers MJCF model design, Gymnasium environment architecture, reward engineering with mathematical formulations, and real training results from the Mesozoic Labs open-source project.Sat, 01 Mar 2025 00:00:00 GMTprojectsmujocomesozoic-labsreinforcement-learningppohttps://www.findingtheta.com/blog/using-reinforcement-learning-for-stock-trading-with-finrl/https://www.findingtheta.com/blog/using-reinforcement-learning-for-stock-trading-with-finrl/A walkthrough of reinforcement learning for stock trading with FinRL: the MDP and Bellman theory underneath it, building a trading environment, and training and backtesting agents with Stable-Baselines3.Sat, 01 Feb 2025 00:00:00 GMTexperimentsreinforcement-learninghttps://www.findingtheta.com/blog/mastering-robotic-manipulation-with-reinforcement-learning-tqc-and-ddpg-for-fetch-environments/https://www.findingtheta.com/blog/mastering-robotic-manipulation-with-reinforcement-learning-tqc-and-ddpg-for-fetch-environments/Using reinforcement learning (RL), specifically Truncated Quantile Critics (TQC) and Deep Deterministic Policy Gradient (DDPG), to solve the Fetch environments in Gymnasium RoboticsWed, 01 Jan 2025 00:00:00 GMTexperimentsgymnasiumroboticshttps://www.findingtheta.com/blog/beginners-guide-to-model-based-reinforcement-learning-mbrl-with-ataris-breakout/https://www.findingtheta.com/blog/beginners-guide-to-model-based-reinforcement-learning-mbrl-with-ataris-breakout/A beginner's walkthrough of Model-Based Reinforcement Learning on Atari's Breakout: training an action-conditional world model that predicts frames, rewards and terminals, then learning a PPO policy entirely inside it with SimPLeSun, 01 Dec 2024 00:00:00 GMTexperimentsatarireinforcement-learninghttps://www.findingtheta.com/blog/using-multi-agent-reinforcement-learning-to-play-openspiels-connect-4-with-rays-rllib/https://www.findingtheta.com/blog/using-multi-agent-reinforcement-learning-to-play-openspiels-connect-4-with-rays-rllib/Discover how self-play can be used to train a reinforcement learning agent to master Connect 4, achieving advanced strategies without human interventionFri, 01 Nov 2024 00:00:00 GMTexperimentsreinforcement-learninghttps://www.findingtheta.com/blog/mastering-ataris-pong-with-reinforcement-learning-overcoming-sparse-rewards-and-optimizing-performance/https://www.findingtheta.com/blog/mastering-ataris-pong-with-reinforcement-learning-overcoming-sparse-rewards-and-optimizing-performance/Train an RL agent to master Atari's Pong with sparse rewards and high-dimensional inputs. Explore preprocessing, replay buffers, and performance-boosting strategiesTue, 01 Oct 2024 00:00:00 GMTexperimentsgymnasiumdqnppoatarihttps://www.findingtheta.com/blog/solving-gymnasiums-car-racing-with-reinforcement-learning/https://www.findingtheta.com/blog/solving-gymnasiums-car-racing-with-reinforcement-learning/Learn how to apply reinforcement learning to solve Gymnasium's Car Racing game, see how different algorithms perform, and explore whether discrete or continuous action spaces are better.Sun, 01 Sep 2024 00:00:00 GMTexperimentsgymnasiumppodqnsachttps://www.findingtheta.com/blog/comparing-how-ppo-sac-and-dqn-perform-on-gymnasiums-lunar-lander/https://www.findingtheta.com/blog/comparing-how-ppo-sac-and-dqn-perform-on-gymnasiums-lunar-lander/Explore how different On-Policy and Off-Policy reinforcement learning algorithms perform on Gymnasium's Lunar LanderThu, 01 Aug 2024 00:00:00 GMTexperimentsdqngymnasiumpposachttps://www.findingtheta.com/blog/solving-gymnasiums-lunar-lander-with-deep-q-learning-dqn/https://www.findingtheta.com/blog/solving-gymnasiums-lunar-lander-with-deep-q-learning-dqn/Learn how the Reinforcement Learning Algorithm Deep Q Learning (DQN) works and apply it to solve Gymnasium's Lunar LanderMon, 01 Jul 2024 00:00:00 GMTexperimentsdqngymnasiumreinforcement-learning