PowerPoint Presentation
Introduction to AI
Francesca Toni (Part I)
Alessandra Russo (Part II)
Course outline
Part I (FT)
• Fundamentals of search and planning in AI
• Resolution and unification and their use in automated reasoning
• Foundations of logic programming
• Rule-based systems for robotics
Part II (AR)
• Foundation of abductive logic programming with algorithms
• (Declarative) Answer Set Programming for real-world applications
• Satisfiability problem and algorithms for Sat-solving
Logistics
• one assessed coursework (pen and paper and Prolog)
• summer term exam (2/2 questions)
• Part I (FT)
• Lectures + tutorials: Tuesdays 16:00-17:00 and Thursdays 9:00-11:00
• Labs(mostly coursework Q&A): Tuesdays 17:00-18:00
• No panopto recordings as they
• hinder spontaneity (in both students and lecturers)
• hinder interactions (amongst students and between lecturers and students
• encourage rote learning
• Yes Piazza
• Slides complemented by reading material – integral part of course
–
Recommended reading: Additional reading:
Chapter 1 Chapter 1
Reading material for the introductory lecture
What is Artificial Intelligence?
like humans rationally
think cognitive (neuro)science logic
act Turing test agents
Design and build systems that
The Turing test
captcha – Completely
Automated Public Turing
Test To Tell Computers and
Humans Apart
6
Cognitive (neuro)science
scientific theories of internal activities of the brain
Predict and test behavior of human subjects
Link to neurological data
Logic – the study of valid arguments
1st meaning: valid chains of reasoning
✓ Every human is mortal, Socrates is human, therefore Socrates is mortal
Every human is mortal, Socrates is human, therefore Socrates is blond
✓ No Martian is human, every human is mortal, therefore some mortal is
not Martian
✓ Axioms of number theory hold, therefore 2 is irrational
2nd meaning: valid disputes
There are two sides to every issue
(Protagoras 490 – 420 BC & Sophists 5th century BC)
Rational agents (Doing the right thing)
What is Artificial Intelligence?
“Intelligent” systems able to interact with /relate to humans
General Data Protection Regulation (GDPR)
“humanisation” of machine-generated decisions
What is Artificial Intelligence?
A collection/combination of methods & systems
• machine learning
• search
• planning
• knowledge representation
• reasoning
• constraint satisfaction
• natural language processing
• robotics and perception
• multiagent systems
27th International Joint Conference on AI
joint with the 23rd European Conference on AI
What is Artificial Intelligence?
• Automated reasoning and inference
• Case-based reasoning
• Cognitive aspects of AI
• Commonsense reasoning
• Constraint processing
• Heuristic search
• High-level computer vision
• Intelligent interfaces
• Intelligent robotics
• Knowledge representation
• Machine learning
• Multiagent systems
• Natural language processing
• Planning and theories of action
• Reasoning under uncertainty or imprecision
How did AI come about?
• “A logical calculus of the ideas immanent in nervous activity”
McCulloch & Pitts 1943:
for any logical expression satisfying certain conditions, one can find a net
behaving in the fashion it describes.
• “Computing Machinery and Intelligence” Turing 1950:
Can machines think?… I believe that at the end of the century the use of words
and general educated opinion will have altered so much that one will be able to
speak of machines thinking without expecting to be contradicted.
• Dartmouth summer research project on AI 1956:
how to make machines use language, form abstractions and concepts, solve
kinds of problems now reserved for humans, and improve themselves.
A partial abridged history of AI
• Newell and Simon (1956): discover proofs in propositional logic
• Wang (1960): prove every theorem in Principia Mathematica
• Samuel (1961): beat humans at checkers
• Minski (1952), Rosenblatt (1958): learning neural networks
• McCarthy, Hayes (1969): logical representations
• Minsky (1975): frames
• Expert systems (1970-80): DENDRAL (chemistry)+MYCIN (medicine)
• Logic-based reasoning (1970-80): Prolog
• Brookes (1987): Intelligence without representation
• Agents and multi-agent systems (1990-)
• NLP: STUDENT (Bobrow 1967 ), SHRDLU (Winograd 1972), CHAT-80 (Warren-Pereira 1982)…IBM Watson
• Semantic web (2001-)
• Deep neural networks (2006-)
What can AI do today?
Answer queries
Play (and win) games
Recognise speech
Recognise faces
Translate across languagesvacuum clean
Drive vehicles
Weak vs Strong AI
•Weak AI: technology leading to tools
•Narrow focus
• Specific domains
•Stong AI (Artificial General Intelligence):
• general-purpose: consciousness? commonsense?
• Ethics
Commonsense
Winograd “schema”
• The council refused demonstrators a permit as they feared violence
• The council refused demonstrators a permit a they advocated violence
• Easy for humans to map they to council and demonstrators respectively
• Grammar/statistical test will not help: commonsense is needed
Introduction – summary
AI
•What (several “definitions”/interpretations)
•What for (several uses/applications)