2. Artificial Intelligence

• Signal Processing Perspective

  • Signal and System

  • Digital Signal Processing

  • Modern Digital Signal Processing

  • “Probabilistic Machine Learning” - a book series by Kevin Murphy

• Machine Learning

  • CMU Intro to ML & lecture videos

  • Berkeley CS189 Intro to ML & lecture videos (Prerequisites: EE16 LinAlg; Math53 Multivariable calculus; CS70 Probability; EE126 Deeper Probability; EE127 Optimization)

  • Waterloo & lecture videos

  • CS229M: Machine Learning Theory

  • Applied Machine Learning 配合CS229

  • CompSciProgram

• Deep Learning

  • Learning map by 李宏毅

  • Dive into deep learning & Notes by 李沐

  • Yann LeCun NYU Deep Learning

  • CS188: Intro to AI for beginners provided by Berkeley

  • CMU MultiModal Machine Learning

• Other Resources

  • Math details by Jeff Miller

  • Open source: Machine Learning with PyTorch and Scikit-Learn

  • Convex Optimization and Approximation from Berkeley EE 227C

  • Playlist: Traditional ML in Python

  • Kaggle

  • 阿里天池

Three Paths to Learn Machine Learning

Path 1 (Bad path)

1. Learn Machine Learning and Deep Learning models only

2. Learn how to use TensorFlow | Keras | Scikit-learn

3. Start using them on the data.

Approximate Learning Time: 6 months to 1 year.

Path 2 (Good path)

1. Learn Linear Algebra, Calculus (differential and integral) and Statistics

2. Learn Machine Learning and Deep Learning

3. Learn how to use TensorFlow | Keras | Scikit-learn

4. Start using them on the data.

Approximate Learning Time: 2 years.

Path 3 (Worst path)

1. Learn pure and abstract math (e.g. Topology, Abstract Algebra, Differential Geometry, Algebraic Geometry)

2. Learn applied maths (e.g. Real analysis, Measure theory, Functional analysis, Topological Degree theory, Partial and Ordinary differential equations, Numerical analysis, Complex Analysis, Optimization)

3. Learn advanced computer science (e.g. Graph theory and combinatorics, Topological data analysis, Advanced algorithms)

4. Learn classical Statistics and probability theory

5. Learn Bayesian Statistics and Bayesian inference theory

6. Learn Machine Learning and Deep Learning technologies

7. Learn how to use TensorFlow | Keras | Scikit-learn

8. Start using them on the data.

Approximate Learning Time: 10 years.