lecture 5

Question 1

Machine Learning

Answer 1

A field of study that gives computers the ability to learn without being explicitly programmed

Question 2

Learning

Answer 2

Learning is the abilty to improve one’s behavior based on experience.

• The range of behaviors is expanded: the agent can do more.
• The accuracy on tasks is improved: the agent can do things better.
• The spped is improved: the agent can do things faster.

Question 3

Components

Answer 3

The following components are part of any learning problem:
*task: the behavior or task that’s being improved: for example: classification, acting in an environment.

*dara: the experience that are being used to improve performance in the task.

*measure of improvment: how can the improvmwnt be measures?
for example: increasing accuracy in prediction, new skills that were not present initially, improves speed.

Question 4

Types of feedback

Answer 4

• Supervised Learning
• Reinforcement Learning
• Unsupervised Learning

Question 5

Supervised learning

Answer 5

he function fis the target function, or concept to be learned• The function his the hypothesis• Fundamental problem of induction:Find a hypothesis which generalizes well, i.e., performs well on unseen examples

Question 6

Inductive bias

Answer 6

reference for/restriction to some hypotheses based on prior knowledge (info not in the training set)• This preference is called inductive bias• Many possible biases – which is right?• Good heuristic isOckham’s razor:prefer the simplesthypothesis consistentwith data

Question 7

Regression

Answer 7

onstruct/adjust hto agree with fon training set•h is consistent if it agrees with fon all examples• E.g., curve fitting (regression task).

Question 8

learning goals: Machine Learning

Answer 8

ell the difference between supervised, unsupervised and reinforcement learning and between regression and classification−
Explain the principles behind Ockham’s razor and how it relates to a model’s inductive bias, its generalization performance and the risk of overfitting
−
Describe different ways to measure the performance of a regressor and classifier