Phase 1: Bootstrapping Fundamentals (Weeks 1-3)
- Goal: Understand how computers "think" so you can write the logic.
- Action:
- Watch CS50 (Weeks 0-4). This solves your "lack algo fundamentals" problem. It explains arrays, loops, and memory visually.
- Read Automate the Boring Stuff. Learn Python syntax without computer science jargon.
- Watch 3Blue1Brown (Neural Networks). Do not calculate anything. Just watch the animations to understand how the "brain" works.
Phase 2: "Top-Down" Machine Learning (Weeks 4-7)
- Goal: Build an image classifier before you understand the math.
- Action:
- Take Fast.ai (Part 1). This course is designed exactly for people like you. You will write code in lesson 1.
- Use Google Colab to run your code so you don't have to install complex drivers on your PC.
Phase 3: Building the "Imitation" Bot (Week 8)
- Goal: Build a bot that copies Grandmasters (Supervised Learning). This is much easier than AlphaZero.
- Action:
- Install python-chess. It handles the rules.
- Download Lichess data.
- Follow Erik Bernhardsson’s guide. Train a Neural Network to guess the move a human played.
Phase 4: The "AlphaZero" Bot (Week 10+)
- Goal: Make the bot teach itself.
- Action:
- Do not read the AlphaZero paper (too much math).
- Read Surag Nair’s "Simple AlphaZero" Article.
- Clone the AlphaZero General repo.
- Modify the "Game" file in that repo to use
python-chess.
Here is every single resource mentioned, categorized. I have marked the ones best for your specific situation with a ⭐.
- ⭐ CS50: Introduction to Computer Science (Harvard/edX) - Best for fixing algo fundamentals.
- ⭐ Automate the Boring Stuff with Python - Best for practical Python.
- FreeCodeCamp: Python for Data Science (YouTube) - Quick syntax.
- Python Data Science Handbook - Reference for NumPy/Pandas.
- ⭐ 3Blue1Brown: Essence of Neural Networks - Watch this first.
- 3Blue1Brown: Essence of Linear Algebra - Essential for understanding matrices.
- 3Blue1Brown: Essence of Calculus - Theory behind "Learning".
- Khan Academy: Linear Algebra - Only if you want to do manual calculations.
- ⭐ Fast.ai: Practical Deep Learning for Coders - Best "Code-First" course. Low math.
- Kaggle Learn - Interactive coding exercises.
- Machine Learning Specialization (Andrew Ng) - Great theory, but math-heavy.
- Google's Machine Learning Crash Course - Fast-paced TensorFlow intro.
- Deep Learning Book (Ian Goodfellow) - Advanced academic textbook.
- Deep Learning with PyTorch: A 60 Minute Blitz - Official PyTorch docs.
- DeepLearning.AI TensorFlow Developer - TensorFlow certification.
- ⭐ Hugging Face Deep RL Course - Practical, library-based.
- OpenAI Spinning Up - Good code snippets, intermediate theory.
- David Silver: RL Lectures (DeepMind) - The "Bible" of RL theory (Math heavy).
- Reinforcement Learning: An Introduction (Sutton & Barto) - The academic textbook.
- ⭐ python-chess Documentation - Use this! Don't write your own rules.
- ⭐ Coding Adventure: Chess AI (Sebastian Lague) - Visual explanation of Search.
- Chess Programming Wiki - Technical encyclopedia.
- Sentdex: Python Chess AI Series - Video log of building an engine.
- John Levine's MCTS Explainer - Explains the search algorithm.
- ⭐ AlphaZero General (Surag Nair - GitHub) - Clean Python code to study.
- ⭐ A Simple AlphaZero Tutorial (Surag Nair - Article) - Explains the code above.
- AlphaZero Paper (DeepMind) - The original math paper.
- Leela Chess Zero (Lc0) - Professional open-source engine.
- Erik Bernhardsson: Deep Learning for Chess - Guide for the "Imitation" approach.
- ⭐ Google Colab - Free GPUs for training.
- Lichess Database - Human game data.
- CCRL (Computer Chess Rating Lists) - Computer game data.
- FICS Games Database - Historical game data.
- Kaggle Kernels - Alternative to Colab.