JP’s ML Corner

While I am helping to develop polaris, gitlab, I want to collect in this blog post all the good links and information I find about machine learning – computer science’s final frontier.

0 Learning about ML

• [1] Machine Learning Mastery – Start here
• [2] Python Data Science Handbook
• [3] Practical Deep Learning for Coders, v3
• [4] Forecasting: Principles and Practice
• [5] Machine Learning for Beginners: An Introduction to Neural Networks
• A Tour of Machine Learning Algorithms – Machine Learning Mastery

1 ML Maps

• Machine Learning Mindmap / Cheatsheet – GitHub – dformoso

2 Algorithm & Documentation

• XGBoost Documentation
• SciKit-Learn Pipelines
  • Documentation
  • User Guide
  • Why you should use scikit-learn’s Pipeline object – signal-to-noise.xyz
• Automated Machine Learning
  • Automated Machine Learning, Open Access Book, Springer
• Context Aware Systems
  • Context Aware Machine Learning, Zeng 2019, arXiv

2.1 Testing

• Pytest
  • Parameterizing Tests (Pytest Doc)

3 Classification & Regression

3.1 Classification

3.2 Clustering

3.3 Filtering

3.4 Ranking

3.5 Regression

• Finding the right fit: Understanding regression in machine learning
  ” It falls under supervised learning wherein the algorithm is trained with both input features and output labels. It helps in establishing a relationship among the variables by estimating how one variable affects the other. “ – Vihar Kurama, builtin.com

4 Feature Engineering

• fets – Feature Extraction Transformers Set fets,gitlab

5 Feature Selection

• Xverse: selection by voting: Introducing Xverse, Medium

6 ML Tools

6.1 MLflow

• MLFlow – Documentation
• 3 concepts MLflow Doc
  • Tracking
    “API and UI for logging parameters, code versions, metrics, and artifacts“
    • run: “executes data science code”
    • experiment: organizes multiple runs
  • Projects
    “standard format for packaging reusable data science code“
  • Models
    “offer a convention for packaging machine learning models in multiple flavors“

7 Reinforcement Learning

7.x Q Learning

7.x Apprenticeship Learning – Imitation Learning

“Some domains are so complex that it is difficult to define a reward function for use in reinforcement learning.”

“seek additional sources of information about the appropriate reward function. One such source is the behavior of agents who are already optimizing (or, let’s say, nearly optimizing) that reward function—in this case, expert human drivers.
The general field of apprenticeship learning studies the process of learning how to behave well given observations of expert behavior.”

“[A]ssuming the environment is observable, we apply supervised learning to the observed state–action pairs to learn a policy 𝜋(s). This is called imitation learning.” 

Artificial Intelligence: A Modern Approach 22.6

99 Testing

• Pytest
  • Effective Python Testing With Pytest – RealPython
    “pytest is a feature-rich, plugin-based ecosystem for testing your Python code.”