Shortcut Resources
1️⃣ Overview — The AI → ML → DL → GenAI Hierarchy
Artificial Intelligence (AI) is the broad concept of making machines mimic human thinking. Machine Learning (ML) uses data to let machines learn automatically. Deep Learning (DL) uses multi-layer neural networks for complex tasks. Generative AI (GenAI) uses transformer models to create — text, images, code, etc.
2️⃣ Learning Roadmap (Visual Mindmaps)
These diagrams summarize AI domains and ML algorithms visually — no need for heavy text. They cover the relationships between ML, DL, NLP, CV, and Reinforcement Learning.
3️⃣ 📚 Learning Resources & Playlists
🐍 Cheatsheets
- Python Cheatsheet → CodeWithMosh PDF
- AI/ML Cheatsheet → GitHub: AI-ML Cheatsheets
▶️ YouTube Playlists
- PyTorch Crash Course → Learn PyTorch - FreeCodeCamp
- Visual Intuition (3Blue1Brown) → YouTube Channel
- AI Concepts in 100 Seconds (Fireship) → Fireship AI Overview
- 5 PyTorch Projects (Hands-on) → Projects Video
- Reinforcement Learning (Practical) → Code Implementation
- Reinforcement Learning (Theory) → Concept Walkthrough
4️⃣ 🧩 Core Machine Learning Algorithm Families
| Category | Quick Understanding (Interview-Style) |
|---|---|
| Regression | Predicts continuous values — e.g., house price, temperature. |
| Classification / Decision Trees | Categorizes things into groups using yes/no questions (e.g., is it spam?). |
| Clustering | Groups similar data points automatically (e.g., customer segmentation). |
| Bayesian Models | Make predictions using probability and prior knowledge (like reasoning with uncertainty). |
| Rule-Based Systems | Use predefined “if-then” logic — simple, interpretable, and fast. |
| Regularization Methods | Prevent overfitting — helps models generalize better to new data. |
| Dimensionality Reduction | Simplifies data by reducing features (like PCA) to focus on what matters most. |
| Ensemble Methods | Combine multiple models (e.g., Random Forest, Boosting) to improve accuracy. |
| Neural Networks | Layers of artificial neurons — core of Deep Learning for vision, NLP, etc. |
| Instance-Based | Store past examples and compare new data to them (like KNN). |