AI Notes
Best resources. Learning path, recommendations. Learning 2024.
Foundational Textbooks
Artificial Intelligence: A Modern Approach (Russell & Norvig)
The definitive AI textbook used in universities worldwide. Covers search, logic, planning, probability, machine learning, NLP, robotics, and philosophy of AI. Now in its 4th edition (2020), updated with modern deep learning and ethical considerations. Best for: comprehensive AI foundation, course reference.
| Search & Planning | Chapters 3-5, 10-11 |
| Machine Learning | Chapters 19-22 |
| NLP & Perception | Chapters 23-26 |
| Robotics | Chapter 26 |
| Philosophy & Ethics | Chapters 27-28 |
Deep Learning (Goodfellow, Bengio, Courville)
The authoritative deep learning textbook covering mathematical foundations (linear algebra, probability, optimization), core architectures (feedforward, CNN, RNN), and advanced topics (autoencoders, GANs, representation learning). Available free online at deeplearningbook.org.
Pattern Recognition and Machine Learning (Bishop)
The gold standard for probabilistic ML. Rigorous Bayesian treatment of regression, classification, neural networks, kernel methods, and graphical models. Mathematics-heavy but rewarding. Best for: graduate-level understanding of ML foundations.
Hands-On Machine Learning (Géron)
Practical guide using scikit-learn, Keras, and TensorFlow. Excellent balance of theory and implementation. Covers end-to-end ML projects from data preparation to deployment. Best for: practitioners who learn by building.
Specialized Books
Natural Language Processing
| Covers | parsing, semantics, dialogue, machine translation |
| Covers | BERT, GPT, T5 for real applications |
| Best for | immediate practical NLP skills |
Reinforcement Learning
Reinforcement Learning: An Introduction (Sutton & Barto):
THE RL bible. Free online (2nd edition).
Covers: bandits, MDP, TD-learning, policy gradient, function approx
Mathematical but accessible. Every RL researcher has read this.Computer Vision
Online Courses
Beginner Level
Andrew Ng's Machine Learning (Coursera/Stanford)
The course that launched ML education online (2012)
Covers: regression, neural networks, SVMs, unsupervised
Updated version (2022): uses Python instead of Octave
Duration: 11 weeks, ~5 hours/week
Fast.ai Practical Deep Learning:
Top-down teaching: build things first, understand later
Covers: vision, NLP, tabular, collaborative filtering
Free, excellent for self-learners
Duration: 7 lessons, each 2 hours
CS50's Introduction to AI with Python (Harvard/edX)
Broad AI survey: search, knowledge, uncertainty, learning, NLP
Hands-on Python projects
Free, self-paced
Intermediate Level
| Courses | NN basics, Optimization, Structuring Projects, CNNs, Sequences |
| Duration | ~4 months at 5 hours/week |
| Best for | Solid DL foundation |
| Stanford CS224N | NLP with Deep Learning: |
| Covers | word vectors, RNNs, attention, transformers, pretraining |
| Best for | serious NLP understanding |
| Stanford CS231N | Convolutional Neural Networks: |
| Covers | image classification, detection, segmentation, generation |
Advanced Level
| Stanford CS229 | Machine Learning (Full): |
| Covers | supervised, unsupervised, RL with proofs |
| Stanford CS330 | Multi-Task and Meta-Learning: |
| Advanced topics | learning to learn, few-shot, transfer |
| Covers | tokenization, fine-tuning, evaluation, deployment |
Research Resources
Staying Current
Papers
arXiv.org: Pre-print server (search cs.AI, cs.CL, cs.CV, cs.LG)
Papers With Code: Papers + implementation + benchmarks
Semantic Scholar: AI-powered paper search and recommendations
Newsletters
The Batch (deeplearning.ai): Weekly AI news curated by Andrew Ng
Import AI: Weekly newsletter by Jack Clark
AI Alignment Forum: Safety-focused research discussion
Conferences (top venues)
NeurIPS: Broad ML/AI (December)
ICML: Machine learning theory and methods (July)
ICLR: Deep learning and representation learning (May)
ACL/EMNLP: Natural language processing
CVPR/ICCV: Computer vision
AAAI/IJCAI: Broad AI
Practice Platforms
| Kaggle | Competitions, datasets, notebooks, courses |
| Start with | Titanic, MNIST, House Prices |
| Progress to | featured competitions, medals |
| LeetCode (ML problems) | Implement algorithms from scratch |
| HackerRank AI track | Search, game theory, statistics |
| Google Colab | Free GPU for experimentation |
Suggested Learning Paths
Path 1: ML Engineer (Industry)
| Month 1-2 | Andrew Ng's ML course + Python proficiency |
| Month 3-4 | Deep Learning Specialization + Kaggle projects |
| Month 5-6 | Hands-On ML book (end-to-end projects) |
| Month 7-8 | Specialization (NLP or CV) + deployment skills |
| Month 9+ | Portfolio projects, interview prep, job applications |
Path 2: Research Scientist (Academic)
| Year 1 | AIMA + Bishop's PRML + linear algebra/probability review |
| Year 2 | Deep Learning book + CS229/CS231N/CS224N + reproduce papers |
| Year 3 | Read cutting-edge papers + contribute to open-source + publish |
Path 3: AI Product Manager
| Month 1-2 | Andrew Ng's AI for Everyone + ML course (high-level) |
| Month 3-4 | Fast.ai (practical intuition) |
| Month 5-6 | Domain-specific AI applications + ethics/policy reading |
Interview Questions
Q: What's the best way to learn AI/ML as a beginner? A: Start with Andrew Ng's Machine Learning course for foundations, then Fast.ai for practical deep learning. Build projects immediately—don't wait until you "know enough." Kaggle competitions provide structured problems. The key is consistent daily practice (even 30 minutes) rather than sporadic cramming.
Q: Should I focus on theory or implementation? A: Both, but weight depends on your goal. Industry roles: 70% implementation, 30% theory. Research roles: 50/50. Start with implementation (build working systems, get intuition), then deepen theory as needed to solve harder problems. Understanding WHY something works helps you fix it when it breaks.
Exam Focus
Revise definitions, diagrams, examples, and short-answer points for Recommended Books & Courses.
Interview Use
Prepare one clear explanation, one practical example, and one common mistake for this Artificial Intelligence topic.
Search Terms
artificial-intelligence, artificial intelligence, artificial, intelligence, resources, recommended, books, and
Related Artificial Intelligence Topics