[ExI] AI milestones
anders at aleph.se
Wed Feb 29 18:40:15 UTC 2012
I have been asked to make a list of milestones in AI development. This
is my current list, what have I missed? And what *actually* matters?
Academic papers easily overstate their importance or rig demos, actual
use is often hidden inside industry (few papers or reports), and
impressiveness might correlate very loosely with real importance.
1950 First serious analysis of whether machines can think. Alan Turing,
Computing Machinery and Intelligence, Mind, 59:433–460, 1950.
1955-1956 Logic Theorist by Allen Newell, Herbert Simon and J. C. Shaw.
Intended to mimic human problem solving skills. It eventually proved 38
of the first 52 theorems in Whitehead and Russell's Principia
Mathematica (with new and more elegant proofs for some). They later
developed the General Problem Solver for more general domains.
McCorduck, Pamela (2004), Machines Who Think (2nd ed.), Natick, MA: A.
K. Peters, Ltd., p. 167
1959 Arthur Samuel’s checkers program. Originally written in 1952, the
1955 version incorporated machine learning. First demonstration of a
program that could learn to play a game better than its creator.
Samuel, Arthur L. (July 1959), "Some studies in machine learning using
the game of checkers", IBM Journal of Research and Development 3 (3):
Schaeffer, Jonathan. One Jump Ahead:: Challenging Human Supremacy in
Checkers, 1997,2009, Springer, ISBN 978-0-387-76575-4. Chapter 6.
1964 ELIZA by Joseph Weizenbaum demonstrates a conversation interface
and that humans are very easily fooled into believing there is
intelligence behind it. Jospeh Weizenbaum, Computer Power and Human
Reason: From Judgment to Calculation (1976)
1964 ANALOGY by Thomas G. Evans solves geometric analogy problems of the
same kind as found in standard intelligence tests. Performance was about
the level of a dull Grade 9 student. Thomas G. Evans, A heuristic
program to solve geometric-analogy problems, AFIPS '64 (Spring)
Proceedings of the April 21-23, 1964, spring joint computer conference,
ACM New York, NY, USA
1975 The Meta-Dendral learning program produced new results in chemistry
(some rules of mass spectrometry) the first scientific discoveries by a
computer to be published in a refereed journal.
1979 The first human killed in an industrial robot accident. $10 Million
Awarded To Family Of U.S. Plant Worker Killed By Robot", Ottawa Citizen,
August 11, 1983, p14
1979 The backgammon program BKG by Hans Berliner defeats the reigning
world champion. This was the first computer program to defeat a world
champion in any game (although Berliner stated that it was due to luck
with the dice rolls).
Berliner, Hans, et al. "Backgammon program beats world champ", ACM
SIGART Bulletin, Issue 69. January 1980. pp 6-9.
1979 MYCIN, an expert system, demonstrates performance within its own
domain (bacterial infection) as good as some experts and better than
general practitioners. Yu, V.L., et al. (1979). "Antimicrobial selection
by a computer: a blinded evaluation by infectious disease experts".
Journal of the American Medical Association 242 (12): 1279–1282. PMID
1979 Cordell Green, David Barstow, Elaine Kant and others at Stanford
demonstrated the CHI system for automatic programming.
1981-1982 Douglas Lenat’s heuristic program Eurisko wins the United
States Traveller TCS championship both years, forcing rule changes due
to its unorthodox strategies. Douglas B. Lenat, Eurisko: A program that
learns new heuristics and domain concepts: The nature of Heuristics III:
Program design and results, Artificial Intelligence, vol 21:1-2, March
1983, p. 61-98
Douglas B. Lenat, Learning program helps win national fleet wargame
tournament, ACM SIGART Bulletin Issue 79, January 1982
Malcolm Gladwell, How David Beats Goliath: When underdogs break the
rules, May 11 2009
1991 The DART tool for automated logistics planning and scheduling was
used during Operation Desert Storm with great success, sometimes
surprising military planners. DARPA claims this single application more
than paid back their 30 year investment in AI. Cross, Stephen E.;
Edward, Walker (1994). Zweben, Monte; Fox, Mark S.. eds. Intelligent
Scheduling. University of Michigan: Morgan Kaufmann. pp. 711–729.
Reese Hedberg, Sarah (2002). "DART: Revolutionizing Logistics Planning".
IEEE Intelligent Systems (IEEE) 17 (3): 81–83
1992 The backgammon program TD-Gammon by Gerry Tesauro reaches
championship-level ability through reinforcement learning and self-play.
Tesauro, Gerald (March 1995). "Temporal Difference Learning and
TD-Gammon". Communications of the ACM 38 (3).
1994 Automatic speech recognition reaches range of human transcription
errors for air travel planning kiosk speech.
1996 Genetic algorithms evolve analog electronic circuits competitive
with human designers. Koza, J.R.; Bennett, F.H., III; Andre, D.; Keane,
M.A.; Four problems for which a computer program evolved by genetic
programming is competitive with human performance, Proceedings of IEEE
International Conference on Evolutionary Computation, 1996.,
20-22 May 1996 Nagoya, pp. 1 - 10
1997 The Deep Blue chess machine beats the world chess champion, Garry
Kasparov.Since then chess computers have increased in power.
Monty Newborn, Beyond Deep Blue; chess in the stratosphere, Springer 2011
1999 The crossword-solving program Proverb better than the average
crossword-solver. Proverb: The probabilistic cruciverbalist. By Greg A.
Keim, Noam Shazeer, Michael L. Littman, Sushant Agarwal, Catherine M.
Cheves, Joseph Fitzgerald, Jason Grosland, Fan Jiang, Shannon Pollard,
and Karl Weinmeister. 1999. In Proceedings of the Sixteenth National
Conference on Artificial Intelligence, 710-717. Menlo Park, Calif.: AAAI
2000 Marcus Hutter’s universal artificial intelligence agent (AIXI), a
theoretical but well-defined agent that behaves optimally in any
computable environment. A computable but still impractical algorithm
(AIXItl) has been implemented. http://www.hutter1.net/ai/paixi.htm
2001 The Unmanned Aerial Vehicle Global Hawk made the first autonomous
non-stop flight over the Pacific Ocean from Edwards Air Force Base in
California to RAAF Base Edinburgh in Southern Australia.
2001 Robotic trading agents consistently beat humans in a commodities
trading game. Robots Beat Humans in Trading Battle. BBC.com (August 8th,
Agent-Human Interactions in the Continuous Double Auction", IBM
T.J.Watson Research Center, August, 2001
2002 Scrabble playing software exceeds best human players. Sheppard, B.
(2002). "World-championship-caliber Scrabble". Artificial Intelligence
134: 241–275. doi:10.1016/S0004-3702(01)00166-7
2004 Automatic speech recognition of broadcast English reaches 10% word
2004-2007 The DARPA Grand Challenge for autonomous vehicles. In 2004
none of the vehicles succeeded the course (the furthest distance
travelled before getting stuck was a little over 11 km). In 2005 five
vehicles successfully navigated the 240 km course (the Stanford vehicle
completed it in 6:54). In 2007 the urban challenge had a 96 km course in
urban terrain with other vehicles and a requirement to follow traffic
rules; the winning vehicle completed it in 4:10.
Sebastian Thrun, Mike Motemerlo, Hendrik Dahlkamp et al., Stanley: The
Robot that Won
the DARPA Grand Challenge, Journal of Robotic Systems - Special Issue on
the DARPA Grand Challenge, Part 2, Volume 23 Issue 9, September 2006
Martin Buehler, Karl Iagnemma, Sanjiv Singh (eds.) The DARPA Urban
Challenge, Springer tracts in advanced robotics 56, 2009
2005-2006 Bridge playing software is on par with the best bridge teams.
2007 The STANDUP pun generator is found consistently entertaining by
children. Graeme Ritchie, Ruli Manurung, Helen Pain, Annalu Waller, Rolf
Black, Dave O'Mara. "A practical application of computational humour."
In Cardoso, A. & Wiggins, G. (Ed.) Proceedings of the 4th. International
Joint Workshop on Computational Creativity, London, UK, 2007, pp. 91-98.
2009 OCR accuracy for commercial OCR software between 71% to 98% for
typeset text. For ISO 1073-1:1976 and similar typefaces intended for OCR
performance is human-equivalent. Holley, Rose (April 2009). "How Good
Can It Get? Analysing and Improving OCR Accuracy in Large Scale Historic
Newspaper Digitisation Programs". D-Lib Magazine.
2010 Demonstration that a surgical robot can learn tasks from human
examples, smooth them and perform them 7-10 times faster, reaching
superhuman performance on these tasks. van den Berg, J.; Miller, S.;
Duckworth, D.; Hu, H.; Wan, A.; Xiao-Yu Fu; Goldberg, K.; Abbeel, P.;
Superhuman performance of surgical tasks by robots using iterative
learning from human-guided demonstrations, 2010 IEEE International
Conference on Robotics and Automation (ICRA), 3-7 May 2010, Anchorage,
AK. P. 2074 - 2081
2010 The May 6 “Flash crash”, where the Dow Jones descended 600 points
only to return to norml after a few minutes. Algorithmic and high
frequency trading were blamed as contributing. High frequency trading
corresponds to more than 73% of US trading by volume, and financial news
are increasingly presented in computer-readable form for the trading
2010 The VisLab Intercontinental Autonomous Challenge, a 13,000 km test
run of autonomous vehicles. Four driverless electric vans successfully
ended the drive from Italy to China, arriving at the Shanghai Expo on 28
October, 2010. It was the first intercontinental trip ever with
autonomous vehicles. http://viac.vislab.it/
2010 A jigsaw-puzzle solver handles 400 piece arbitrary image puzzles.
2011 Sentiment analysis for online text at the same level at human
reliability. Andrew L Maas, Raymond E Daly, Peter T Pham, Dan Huang,
Andrew Y Ng, Christopher Potts, Learning Word Vectors for Sentiment Analysis
Computational Linguistics (2011) Volume: 31, Issue: Jne 19-24, 2011,
Publisher: Association for Computational Linguistics, Pages: 142-150
2011 Heuristics evolved using genetic algorithms produce a solver for
the NP-hard solitaire game FreeCell, able to beat high-ranking human
players. Achiya Elyasaf, Ami Hauptmann, Moshe Sipper, GA-FreeCell:
Evolving Solvers for the Game of FreeCell, GECCO’11, July 12–16, 2011,
2011 IBM’s Watson defeats the two greatest Jeopardy! Champions during an
2011 Computer poker players remain sub-human for full ring Texas hold
'em but approaching strong super-human in simpler versions of poker.
Rubin, Jonathan; Watson, Ian (2011). "Computer poker: A review".
Artificial Intelligence. doi:10.1016/j.artint.2010.12.005.
2012 The Zen series of go-playing programs reaches rank 4-5 dan (strong
Future of Humanity Institute
Philosophy Faculty of Oxford University
More information about the extropy-chat