getting-started-machine-learning-archive.md

title	layout
Getting Started in Machine Learning	default

Getting Started in Machine Learning

This page is no longer maintained.

1. Build the breadth of ML knowledge by taking a FULL course in ML, ideally a graduate level at STAT/MATH/CS departments. If not available, then take a complete online course (as described in this guide).
2. Never rely solely on notebooks or online tutorials to learn the basics to avoid the common knowledge gaps. Think of a researcher with zero statistical knowledge trying to learn just from tutorials, surely their published analysis will have serious flaws.

Recommended Book:

Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems 2nd Edition, by Aurélien Géron , and its notebooks.

Other well-known references for statistical background could be used to understand the statistical theories and proofs If needed.

Fundamental background (prerequisites):

• Linear Algebra
• Calculus
• Probabilities and statistics
• Basic programming

General Machine Learning:

For this section, you can take a Machine learning or data science university course, take an online course, or study a textbook. Here we suggest the online track if university courses are not available.

1. For Fundamentals

• If you prefer courses, the gold-standard is: Machine Learning, by Andrew Ng, Stanford

  • Must take for ML basics
  • Viewed by more than 8M students, 2.7M enrolled in the new system
  • Theory, language-independent in general (but examples in Octave, a Matlab-like language)

• If you prefer Textbooks, the free book: An Introduction to Statistical Learning

1. Machine Learning A-Z™: Hands-On Python & R In Data Science

• Udemy, ~$20
• 500k+ students took the course (4.5 stars rating)
• 41 hrs, hands-on projects

1. Theory and practical Python for Data Science and Machine Learning Bootcamp

• Udemy, ~$20
• 21 hrs
• Practical -- Excellent instructor
• Not deep in fundamentals, but good hands-on

Neural Networks and Deep learning:

1. Deep Learning for Molecules and Materials

This textbooks covers deep learning for molecules and materials with code examples and theory. It starts from ML and works up to modern methods in deep learning

2. Deep Learning A-Z™: Hands-On Artificial Neural Networks

• Udemy
• 22 hrs
• Practical, explains the basics,
• but NO mathematics

3. Neural Networks and Deep Learning

• Coursera (free without certification)
• From Andrew Ng Deep.ai
• Theory and how to use → experience
• Can be taken for free (if Audit), or pay to get a certificate
• Part of a full 5 courses specialization (RNN,hyperparameters opt, ...etc)

4. Improving Deep Neural Networks: Hyperparameter tuning, Regularization, and Optimization

• Part of the above specialization

Very comprehensive articles: (not to replace courses)

• Fundamental Techniques of Feature Engineering for Machine Learning (excellent)
• Comparative Study on Classic Machine learning Algorithms
• ML Notebooks
• Which Machine Learning Algorithm Should You Use By Problem Type
• A guide to an efficient way to build neural network architectures- Part I: Hyper-parameter selection and tuning for Dense Networks using Hyper as on Fashion-MNIST
• Google, Rules of Machine Learning
• Google, Machine Learning Crush Course
• Tinker With a Neural Network Right Here in Your Browser
• Comparative Study on Classic Machine learning Algorithms
• Tutorials with ML use in QM