WEEK 2.9

Question 1

data science

Answer 1

activity and field of education

Question 2

activity

Answer 2

data collection/interpretation/analysis via automated and manual methods to posit/evaluate models

Question 3

education

Answer 3

the study of

• Data collection/interpretation/analysis
• Statistical/machine-learning methods
• Model selection

Question 4

What do data science models do?

Answer 4

• codify knowledge in a given area
• generate specific decisions under specific conditions

Question 5

machine-learning algorithms

Answer 5

• very good at detecting patterns
• use well-known statistical techniques to achieve results

Question 6

AI as an activity

Answer 6

ability of an artificial unit to perform intelligent tasks in a (semi) automated fashion

Question 7

AI as an education

Answer 7

study of artifical systems in areas such as natural language processing, reasoning, sensors and robotics

Question 8

weak AI

Answer 8

• artificial units that perform well in certain areas
• e.g. playing games, solving equations, offering diagnoses

Question 9

strong AI

Answer 9

• artificial units that perform well in all alreas and may even be said to have consciousness
• holy grail is to create AI that surpasses human intelligence

Question 10

GOFAI

Answer 10

• artificially intelligent systems focused on rule-like manipulations of symbols
• mimics math and logic

Question 11

expert systems

Answer 11

• developed to mimic decisions taken by human experts
• two components:
  
  • inference engine (logic and heuristics)
  • knowledge base

Question 12

networks

Answer 12

• mimic the interconnected web of neurons that exist in the brain
• most widely used approach in AI

Question 13

computational theory of mind

Answer 13

mind taken to behave like computer

Question 14

machine learning

Answer 14

• sub-branch of AI where algorithm learns as much as possible on own
• learning done from data not experience
• mostly conducted by neural networks

Question 15

connectionists

Answer 15

advocate the use of neural networks to understand brain

Question 16

supervised learning

Answer 16

• algorithm trained with labelled data and then its output is checked in test phases
• we know desired result, aim is to produce correct labels
• useful in classification and curve-fitting tasks

Question 17

unsupervised learning

Answer 17

• algorithm trained but data not labelled
• we don’t know desired result, aim is for the algorithm to produce its own pattern from the data
• useful in clustering tasks