Week 8: Recap of Learning

Question 1

For our brain to produce ‘adaptive’ behaviour it must be capable

Answer 1

of learning!

Question 2

Unsupervised learning

Answer 2

The network learns useful weights soley based on the statistics of input patterns

Question 3

Example of unsupervised learning (3)

Answer 3

Classifying input patterns vis competitive learning (winner takes all)
BCM rule (unlearning below threshold)
Basic Hebbian learning rule

Question 4

Supervised learning (2)

Answer 4

The network is given a desired output to compare its own output to

The difference instructs the change in synaptic weights

Question 5

Examples of supervised learning

Answer 5

Perceptron (i.e., delta rule) , deep learning models,

Question 6

Reinforcement learning

Answer 6

The network is given intermittent feedback in form of punishment and rewards and uses this as a guide to change weights.

Question 7

Supervised learning rule and reinforcement learning rule can be

Answer 7

combined

Question 8

What are the three learning rules? (3)

Answer 8

1. Unsupervised learning
2. Supervised learning
3. Reinforcement learning

Question 9

Example of supervised learning rule

Answer 9

Delta rule

Question 10

There is no one correct learning rule to understand the brain as

Answer 10

depend on the brain area, the type of information that is learned, thestage of development of the organism etc.

Question 11

As the brain develops, in part genetically hard-wired, different areas of the brain become plastic (some stay plastic, some are so only for a brief time window), and experiences (i.e., incoming neural patterns) are fed into different brain areas where plasticity acts

Answer 11

Different parts of the brain may implement different learning rules at different times.