History and Synopsis of AI Flashcards
AI Application Areas
Vision : object, colour, scene, facial
Game playing : chess, poker
Speech recognition
Theorem Proving : four colour theorem
Definition of an AI
The science of understanding intelligent entities and the engineering of intelligent entities. It is a system that thinks that humans and acts like humans with rational thinking and acting.
Rational thinking
Idealised Reasoning
Cognitive Science
To understand and build intelligent machines to stimulate human thinking.
Computer Science
To build intelligent machines to simulate the actions of humans.
Turing Test
1950 test to see if a human could decipher if it was a human or machine talking back to him. Problem was that it could be defeated by cheap tricks, not intelligence.
McCulloch and Pitt’s 1943 Theory
Used Alan Turing’s notion of computation.
Minskey and Edmonds
SNARC
Shannon Turing
Developed the first chess playing programs
Dartmouth Conference
10 attendees, logic theorist program was rejected by Journal of Symbolic Logic.
John McCarthy created name ‘AI’
1960s
Systems in these times almost always failed on more difficult problems.
Combinatorial Explosion
It is the complexity in which programs grow, with different combinations in chess and a rubik’s cube growing drastically with every move, causing problems and led to support being ended towards AI research.
The Lighthill Report on AI in 1973 came to the conclusion that AI researchers had failed to address this problem and was pessimistic regarding research general robotics and natural language processing.
Expert Systems
Tailed towards a specific domain (MYCIN, PROSPECTOR)
Explainable AI
Current challenge because some do not explain the outcome and no-one is able to come to a conclusion on it.
