Success according to human standards, Systems that act like humans, AI is the art of creating machines that perform functions that require intelligence when performed by humans, Methodology: Take an intellectual task at which people are better and make a computer do it, Example: Turing test, Lecture 1a, Slide 16, Systems that think like humans, Today's AI varies greatly from human thinking, Not very relevant.
Success according to an ideal concept of intelligence: rationality., Systems that think rationally, Aristotle: What are ‘correct’ argument and thought processes?, Correctness depends on irrefutability of reasoning processes., Based on logic, Problems:, Not all intelligence is based on logic, Humans don't think rationally, The real world often isn't rational., Systems that act rationally, "Doing the right thing", Maximize goal achievement given available information., Hence requires a defined goal, May not have all information, but should still be able to make the best decision based on available information, Can involve thinking, Can involve action without thinking -reflexes etc., Problems:, Rationality is only applicable in ideal evironments, Rationality is not a good model of reality
Core Aims (1a, slide 6)
Definition: What is intelligence (1a, slide 13)
Summary:, Different people think of AI differently., Two important questions to ask are: Are you concerned with thinking or behaviour? Do you want to model humans or work on an ideal standard?, In this course, we adopt the view that intelligence is concerned mainly with rational action., Ideally, an intelligent agent takes the best possible action in a situation. We will study the problem of building agents that are intelligent in this sense.
Continues on from previous lecture slides (L1b)
Lecturer(s), Dr. Bao Vo, firstname.lastname@example.org, Work Phone 92144756, Office Location EN 504, MAY BE SWITCHING TO ANOTHER OFFICE, Consultation Times:, Monday 15:30 - 17.00
Tutor(s), Minyi Li, email@example.com, Work Phone 92148775
Artificial Intelligence: A Modern Approach, Russell, S.J., Norvig, P., 3rd edition, Prentice-Hall, 2010. [AIMA]
GROUP ASSIGNMENT (Pair is allowed)
Marking, 50% for code - Java, C or C++ (python, ruby on request), 50% for self assessment/peer assessment, associated report
Must Score at least 50%
History, Not Covered In Exam!!!
Aka - Intelligent Agents, Basically just another name for AI
Agents include human, robots, softbots, thermostats, etc.
An agent can perceive its own actions, but not always it effects
Agent Function, If agent perceives something, do something, Lecture 1b example (slide 5,6), Not related to implementing logic - hence last 2 functions on slide 6
Agent Program, The logic used to achieve the agent functions.
Rationality, Rationality != omniscience, Doesn't know the actual outcome of it's action., Rationality != perfection, for on expected performance. perfection is based on actual performance., Information Gathering, Learn form percepts, Agent autonomy
Must Be specified to define a rational agent
PEAS, Performance, Environment, Actuators, Sensors