History of Prolog
- 1965. Alan Robinson published "A Machine-Oriented Logic Based
on the Resolution Principle".
- 1972. Alain Colmerauer and Phillipe Roussel (University of Aix-Marseille)
developed Prolog in collaboration with Robert Kowalski
(University of Edinburgh).
- Prolog is an abbreviation for "Programmation en Logique"
or "Programming in Logic".
- 1974. David Warren wrote an efficient implementation.
- Heavily used in Europe.
- Heavily used in Japan. For example, Prolog served as the
basis for the ICOT Fifth Generation Computer Systems Initiative.
Japan's Fifth Generation Project
- 1982 - 1992.
- Historically, Japan had followed the US and British leads
in computer advancement.
- The goal was to produce an "epoch-making computer" that would
use Prolog to create a desktop system with supercomputer-like
performance and use AI technologies.
- Why the name? The first generation used vacuum tubes.
The second generation used transistors and diodes.
The third generation used integrated circuits.
The fourth generation used microprocessors.
- Goal: Inference computer technologies for knowledge processing.
- Goal: Computer technologies to process large-scale databases
and knowledge bases.
- Goal: High performance workstations.
- Goal: Distributed functional computer technologies.
- Goal: Supercomputers for scientific calculation.
- Goal: 100M to 1G LIPS. Logical Inferences per Second.
- ICOT (Institute for Computer Technology) was formed to spearhead
the project.
- Unfortunately, Prolog did not support concurrency easily.
- Unfortunately, single CPU technology continued to evolve rapidly.
- Unfortunately, software continued to evolve rapidly. For example,
Apple introduced GUIs to the masses.
- Unfortunately, the internet evolved, making large locally-stored
databases unnecessary.
- Unfortunately, AI turned out to be harder than expected.
- After 50 billion yen, the project was terminated and deemed
an abject failure!
Chapter 26.1 - 26.2
Philosophy
- Weak AI Hypothesis: machines could possibly act intelligently
- Strong AI Hypothesis: machines can actually think
Weak AI Counterarguments
- Disability argument: a machine can never do X
- Mathematical argument: Gödel's incompleteness theorem
- Informality argument: human behavior is too complex to capture
by a set of rules
Strong AI Counterarguments
- Consciousness
- Free Will
- Brain-in-a-vat
- Brain prosthesis experiment
- Searle's Chinese Room
- Phenomenology: study of direct experience
- Intentionality: mapping of machine's beliefs to real world
General Questions
- What is intelligence?
- What is a good definition of artificial intelligence?
- What is thinking?
- What is understanding?
- What is creativity?
- What is consciousness?
- What are emotions?