Who are we?

• What is your name?
• Where do you come from? Place and first degree.
• What do you expect to learn?

Tools

• Wiki
• BlackBoard
• GodinGame
• (OneNote on request)

Sessions

• Briefing 10.15-11.15
• Break
• Exercise 11.30-13.30
  • NB Make sure to take a Lunch break
• Status 13.30-14.00
• Exercise 14.00-15.15
• Break
• Debrief 15.30-16.00

This may vary from week to week.

Practical Information

• See Wiki
• Compulsory Assignments
• Learning Outcomes
• Exam Format
• Talk to me!
• Feedback in Class
• Homework

Artificial Intelligence

Debate

1. What is intelligence?
2. What is AI?
3. Is AI possible?

Briefing

• Reading Russel & Norvig Chapter 1
• Lecture Notes

(Artificial?) Intelligence

Philosophy

• How do we know?
  • How do you find your solution?
  • How does the computer find its solution?
• How can we be sure that we know?
  • How do you know if your solution is correct?
• Hume's Is-ought problem
  • How do you know what is?
  • How do you know what ought to be?

Six Areas of AI

1. Natural Language Processing
2. Knowledge Representation
3. Automated Reasoning
4. Machine Learning
5. Computer Vision
6. Robotics

The Analytical Engine

By Alfred Edward Chalon, Science Museum Group, Public Domain

By anonymous

Alan Turing

By unknown.Public Domain

CC BY-SA 3.0

1950s

• Enthusiasm and Expectation 1952-59
• Birth of AI as a separate field 1956
• Simon & Newell: Logic Theorist
• Coincides with Scientification
• Lisp 1958, Eliza, GPL
• Machine Evolution 1958

Eliza

Link

By Ulrich Hansen, Germany (Journalist), CC BY-SA 3.0,Wikimedia Commons

Reality 1966-73

• optimistic ten-year predictions fail
• three challenges
  1. AI has no domain knowledge
  2. intractible problems
  3. limited representations (two perceptrons do not suffice)

Knowledge-based systems 1969-79

• Previous limitation: weak methods which do not scale. Brute force?
• Solution: knowledge bases

Early 1980s

• Industrilisation
• DEC has the first commercial expert system 1982

Mid-1980s

• Scientific methods
• Machine Learning is Statistics
• Neural Networks return

1995

• Intelligent Agents
• Rebirth of the whole agent vision

Turn of the Millenium

• Big Data
• Exploding Storage Capacity
• the Internet grows

2011

• Computing Power
  • Vector processors
• Deep learning
• Reinforcement learning

Applications

• Autonomous Ships (Ottar, Robin)
  • And flying drones (Erlend)
• Medical Image Processing (Hans Georg)
  • Diagnosis, e.g. cancer detection
  • Image segmentation to make 3D models of the patient
• EEG (Brain Waves) (Robin)
• Cosmology - detect dark matter (Hans Georg, Ben David)
• Information Security (previously Hans Georg)
  • detect clandestine communication (steganalysis)
  • intrusion detection

Is it all good?

By George R. Caron

Semester at a glance

1. Programming and Problem-Solving ($\sim 6$ weeks)
2. Ethics: When it helps, when it hurts, and when it fails (long and thin)
3. Genetic Algorithms (2-3 weeks)
4. Machine Learning, Statistics, and Data Analysis (1-2 weeks)
5. Reinforcement Learning (2-3 weeks)

Programming and Problem Solving

• Python
  • You may use any language, but I use python for (most) examples.
• GodinGame
  • Suggested Problems
• In-house skeleton code published on github
  • AI
  • Genetic Algorithms

Problems to be Solved

Make an intelligent agent.

Your program has to

1. Look at the problem.
2. Find a good solution.
3. Act accordingly.
4. (in some games) iterate