Types of Machine Learning problems: Supervised, Unsupervised, and Reinforcement Learning

Created by Lily Vittayarukskul for SVAI research community. Open to collaborators!

Introduction

Supervised Learning

• we have a set of features or inputs X (for example, an image) and our model will predict a target or output variable y

• Types:

  • Classification (e.g. spam/not spam), Regression (e.g. housing prices)

Unsupervised Learning

• Identify patterns in the data

• Types:

  • Clustering: inherent groupings in the input

  • Association: discover rules that describe portions of the input data

Reinforcement Learning

• the input itself depends on actions we take

• robot is known as an agent, and is in some environment (surrounding). At each time step, it can take some action and it might receive some reward.

Hybrid Models

• Semi-supervised learning: subset of data is not labelled (e.g. image recognition)

• Transfer Learning: transferring our knowledge/model from one domain to another

  • E.g. learn a model on the alphabets, and re-use the model in some way when learning a model on the sentences.

Note: A majority of the successful machine learning products currently fall under the category of supervised learning.

Important Note

Unsupervised and reinforcement learning are areas of active research, and we've recently made significant progress in both with algorithms such as Generative Adversarial Networks (for unsupervised learning) and Deep Q-networks (for reinforcement learning).