OUTLINE.md

Machine Learning in Javascript

Steve Purves, Expero inc.

Course Outline

• Icebreaker
  • Warm Up
  • Menti Meter
• Introduction
  • Why not JavaScript? (slides)
    • Where ML in JS takes us (slides)
    • What the ML JS landscape is like
  • Course Outline (md)
• Checking Environment Setup & Sharpening of Tools
  • Installation Instructions
  • Docker Option
  • What's Loaded (nb)
  • [Hello Notebook (nb)]0.1_hello_notebook.ipynb)
  • Hello Plotly
  • Hello Datasets
• Part 1 - Key Concepts
  • Vectors & Spaces (nb)
  • Distances & Costs (nb)
  • Exercise: Applying a distance measure
  • Classification (slides)
  • Error, Cost, Loss & Learning (nb)
  • Measuring Success (nb)
    • training/testing
    • cross validaton
    • scoring
    • confusion matrix
  • Fitting a line (nb)
  • Summarising (slides)
• Part 2 - Classical Approaches
  • Unsupervised
    • [->] K-means Clustering (ml.js)
    • Determining Number of Clusers (ml.js)
    • Emsemble Methods (slides)
    • Stretch:
      • Gaussian Mixure Models (gmm.js)
      • Heirarchical Learning (ml.js)
  • Embeddings
    • Principal Component Analysis (ml.js)
    • tSNE
  • Supervised Learning
    • K-Nearest Neighbor (ml.js)
    • Support Vector Machines (ml.js)
• Part 3 - Neural Networks & Deep Learning
  • Code your own neuron (nb)
  • Fully Connected Network (nb)
  • Convolutional Neural Network (nb)
• Part 4 - Running large models in the browser with Keras.js
  • Running Inception

References

Sources of Data

• UCI Machine Learning Repo

  • Wine Dataset
  • Acknowledgement - Lichman, M. (2013). UCI Machine Learning Repository. Irvine, CA: University of California, School of Information and Computer Science.

• Big Lists

  • https://github.com/viisar/awesome-datasets
  • https://github.com/caesar0301/awesome-public-datasets
  • https://github.com/caesar0301/awesome-public-datasets#machine-learning