Chapter 1: Machine Learning Basics Flashcards
Define machine learning
Machine learns a task using a performance metric and experience
What is a query
Something the machine has never seen before
What are the key stages of machine learning
- Construct a model
- Training
- Evaluation
What is optimisation
Find the minimum of a real valued function
What is supervised learning
There is a target output for each data pattern. This pair is called a training example
What does a supervised learning model learn
The relationship between the data pattern and the target output
Give examples of where supervised learning is used
Classification
Regression
Ordinal regression
What is unsupervised learning? What does the system learn
There is no explicit teacher.
The system forms and understanding of hidden structure
Give examples of unsupervised learning
Clustering
Generative modeling
Unsupervised representation learning
What is reinforcement learning
A supervised learning technique where the type of supervision is different. The aim is to get a reward rather than optimise
What are the three ingredients in the ML pipeline
Construct the model function
Construct the loss function
Optimise the loss function
What are the two approaches to constructing a model
Model the target
Model the posterior
What are some methods to construct a mod by modelling the target
Linear model
Linear basis function
Kernel method
Neural network
What are some methods for constructing a model by modelling the posterior
Logic regression
Bayesian regression
Naive bayes
What are the different loss functions
Sum of squares
Hinge loss
Cross entropy
Mean squared
Likelihood
Log likelihood
