CIT 6261:
Advanced Artificial Intelligence
Introduction
Lecturer: Dr. Md. Saiful Islam
Course Overview
• Advanced search techniques in AI
• Knowledge based system design
• Advanced plan generation systems
• Bayesian network and probabilistic reasoning
• Learning in neural belief networks
• Practical natural language processing
• Computer vision
• Introduction to robotics
The field of Artificial Intelligence
• Goal: To build autonomous intelligent entities
which think and act rationally like human beings.
• Computes with human-level intelligence
What is AI?
• [The automation of] activities • The study of mental faculties
that we associate with human through the use of
thinking, activities such as computational models
decision-making, problem -- Charniak + McDermott, 1985
solving, learning …
-- Bellman, 1978
• The branch of computer
• The study of how to make science that is concerned with
computers do things at which, the automation of intelligent
at the moment, people are behavior
better -- Luger + Stubblefield, 1993
-- Rich + Knight, 1991
What is AI? .. continued
• Systems that think • Systems that think
like humans rationally
• Systems that act like • Systems that act
humans rationally.
Acting humanly: Turing Test
Turing(1950) “Computing machinery and intelligence”
• “Can machines think?” → “Can machines behave intelligently?”
• A test of an intelligent machine which is indistinguishable from
human.
• Operational test for intelligent behavior: the Limitation Game
• Predicted by 2000, a machine might have a 30% chance of fooling
a lay person for 5 minutes.
• Anticipated all major arguments against AI in following 50 years.
• Suggested major areas of AI: knowledge representation,
automated reasoning, natural language processing, machine
learning + computer vision, robotics.
Thinking humanly: Cognitive Science approach
• Needs Theory of Human mind → express the theory as a computer
program.
• If program’s input, output and timing behaviors match corresponding
human behaviors, we can say the program has some intelligence.
• Cognitive Science:
– 1960s “cognitive revolution”: information-processing psychology replaced
prevailing orthodox of behaviorism
– computer models from AI + experimental techniques from Psychology
– Aim: Construct precise/testable theories of the workings of the human mind.
• Requires scientific theories of internal activities of the brain
1) Predicting and testing behavior of human subjects (top-down)
Or 2) Direct identification from neurological data (bottom-up)
• Both approaches (roughly, Cognitive Science and Cognitive
Neuroscience) are now distinct from AI.
• Both share with AI the following characteristic:
the available theories do not explain (or engender) anything resembling
human-level general intelligence
Thinking rationally: Laws of Thought
• Aristotle: what are correct arguments/thought process?
– Syllogism (A is B and B is C, therefore A is C)
• Several Greek schools developed various forms of logic:
notation and rules of derivation for thought.
• Directs line through mathematics and philosophy to modern AI.
– Formal logic (19th –early 20th C.) : syllogism ⇒ formal logic
– Formal logic provided a precise notation for statements about all kinds of
things and their relations.
– Logicist tradition: build an intelligent systems through programs
which take a description of a problem in logical
notation and find the solution to the problem,
if one exists.
• Aim: Build computational reasoning system through the formal logic.
• Problems:
1) Not all intelligent behavior is medicated by logical deliberation
(e.g., when knowledge is not 100% certain).
2) Problem with few dozen of facts can exhaust the computational resources of
any computer.
Acting rationally
• Rational behavior: doing the right thing
• The right thing: that which is expected to maximize goal
achievement given the available information
⇒ Rational agents
• rational reasoning (thinking) ⎯→ rational action (behavior),
however,
rational reasoning ←/⎯ rational action.
- (e.g. recoiling form hot stove)
• Aristotle (Nicomachean Ethics);
Every art and every inquiry, and similarly every action
and pursuit, is thought to aim at some good.
• Advantage:
– More general
– More amenable to scientific development
Rational agents
• An agent is an entity that perceives and acts
• Rational agent: one that acts so as to achieve one’s goals at
the best, given one’s belief.
• Aim of AI: to study /design rational agents
• Abstractly, an agent is a function from percept histories to
actions:
f: P* → A
• For any given class of environments and tasks, we seek the
agent (or class of agents) with the best performance
• Warning: computational limitations make perfect rationality
unachievable → design best program for given machine
resources.
AI vs. CI
• Artificial Intelligence (AI):
– is based on symbolic representation of knowledge;
– creates expert systems that help to reason;
– “knowledge engineering” is its most important branch.
Computational Intelligence (CI)
– may help to discover knowledge hidden in data.
Review of AI
• Intelligent Agents: Goal-based agent, utility-based agent, learning
agent, mobile agent, ...
• Search: DFS, BFS, Best-first search, A* algorithm, Beam search,
Hill-climbing.
• Advanced search: Iterative-deepening A*, Simulated annealing,
Genetic algorithms for search, Constraint satisfaction search, etc.
• Knowledge representation and Automated reasoning:
– Propositional/Predicate logic, first order logic
– Inference: Resolution, unification, forward/backward chaining
– Rule: facts, rules, meta-rules.
Continued…
• Planning: planning as search, situation calculus, etc.
• Machine Learning
– Inductive learning, Decision-Tree induction, The Nearest neighbor
algorithm, Supervised learning, Unsupervised learning,
Reinforcement learning
– Statistical learning
– Neural network
• Natural language processing, Machine vision, etc.
