On Machine Learning: My Study Plan

↑ Engineering

Anton’s Machine Learning Curriculum Plan

I am comfortable with Kotlin, Java, and Python, is diving into machine learning. I have experience with Minecraft mods, GoDot, Lua, and block programming. Currently, I’ve independently dabbled in the basics of linear regression using Google’s machine learning resources. Here’s my plan for the next steps to solidify foundation in ML while keeping the journey engaging.

Summary:

1. Reinforce Linear Regression

   • Use real-world datasets (e.g., housing prices, web activity) for hands-on practice.

   • Visualize data relationships using tools like matplotlib and seaborn.

2. Introduce Logistic Regression

   • Explain binary classification with logistic regression using real-world data (e.g., Titanic survival data).

   • Visualize the sigmoid function and interpret model predictions using metrics like ROC curves.

3. Data Visualization and Feature Engineering

   • Build data visualization skills using libraries like Pandas and Seaborn.

   • Practice feature engineering to understand the impact of features on model performance.

4. Early Neural Networks

   • Transition into neural networks with simple image classification (e.g., MNIST dataset).

   • Visualize how neurons work in layers and explore TensorFlow Playground.

5. Real Projects and Reflection

   • Implement small end-to-end projects like spam detection or image classification.

   • Use Anton’s interests in gaming and storytelling to frame these projects.

     Resources

     • Kaggle Datasets

     • TensorFlow Playground

     • Fast.ai

     Inspiration

     • ML Day Tokyo