Machine Learning

Concepts

• In Machine Learning, you receive rules based on data and answers provided

• In sequential ML models rules flow straight out of data

• loss function (Machine Learning)

• optimizer function (Machine Learning)

• normalisation

• shuffling

• Dense Layer (ML)

Practical

• Steps to use a ML model
• Steps to use a ML model in Tensorflow Python
• Steps to use a ML model in TensorflowJS

Tools/Entities

• Keras Python library
• TensorflowJS
• Copilot

Glossary to learn

• neutral network: just a model?
• input shape

Resources

• Data Science: Machine Learning | edX

• Neural Networks In JavaScript - Brain.js Tutorial | Scrimba

• You can build a neural network in JavaScript even if you don’t really understand neural networks | by Daniel Simmons | ITNEXT

• Supervised Machine Learning: Regression and Classification | Coursera

• deeplearn - npm

• Basics of TensorFlow for JavaScript development

• Embedded Agency - AI Alignment Forum

• Iterated Amplification - AI Alignment Forum

• Value Learning - AI Alignment Forum

• AGI safety from first principles - AI Alignment Forum

• AI Alignment Curriculum — AGI Safety Fundamentals

Prerequisites (according to someone on the internet)

You’ll need to understand these things before beginning with TF or any other library:

1. Few statistics like std dev , median , mode , diff types of distributions.
2. Numpy arrays and how they work & operations on them.
3. Matrix , their operations , dot products
4. Scalars , vectors , Tensors
5. What’s a CPU , GPU & A TPU.