Marvin Minsky

from Wikipedia, the free encyclopedia
Marvin Minsky (2008)
3D surface profile of a 1 euro coin measured with a confocal microscope (detail)

Marvin Lee Minsky (born August 9, 1927 in New York , † January 24, 2016 in Boston , Massachusetts ) was an American researcher in the field of artificial intelligence (AI). Together with John McCarthy , Nathaniel Rochester and Claude Shannon , he founded the term artificial intelligence at the Dartmouth Conference in 1956 . Later he and Seymour Papert also founded the Artificial Intelligence Laboratory at the Massachusetts Institute of Technology (AI Lab).

He published numerous texts on this subject as well as on related topics of philosophy and also made some inventions. He is considered to be the inventor of the measuring principle later realized in the confocal microscope (1957). Further inventions are mechanical hands and other parts for robots, the Muse - synthesizer for musical variations together with Edward Fredkin and the first logo turtle ( turtle graphic ) together with Seymour Papert. In 1951 he built with Dean Edmonds SNARC (Stochastic Neural Analog Reinforcement Calculator), a neural network computer that simulated the behavior of a mouse in a maze.


His parents were the ophthalmologist Henry Minsky at Mount Sinai Hospital and the Zionist activist Fannie, a daughter of the Austrian Abraham Reiser. Minsky's older sister, Charlotte, became an architect and painter, while his younger sister Ruth became a genetic counselor.

In 1952 he married the doctor Gloria Rudisch. He had two daughters and a son.


Marvin Minsky attended Fieldston School and the Bronx High School of Science in New York. He later studied at Phillips Academy in Andover , Massachusetts. He did his military service in the US Navy from 1944 to 1945 . He obtained a bachelor's degree in mathematics from Harvard University in 1950 and received his doctorate in 1954 from Princeton University under Albert William Tucker ( Theory of Neural-Analog Reinforcement Systems and Its Application to the Brain Model Problem ).

Minsky had been at the Massachusetts Institute of Technology since 1958 ; there he researched and taught until his death. In 1959 he founded an AI working group there with John McCarthy , the Lisp inventor. The group, to which many hackers (such as Richard Stallman ) belonged in the 1970s , was assigned to the MAC project (for Project Mathematics and Computation , later also for Multiple Access Computers) from 1963 , which was only established under the direction of Robert Fano (up to 1968) and then by JCR Licklider (until 1971). In the 1960s there was also a lot of research into models of neural networks, perceptrons (the title of a book by Minsky and Papert). The discovery of some fundamental flaws in such simple neural networks by Minsky and Papert led to the fact that research in this area came to a complete standstill in the late 1960s and was not revived until the 1980s. In addition to AI research (for example in the areas of visual perception, robotics, language), a time-sharing computer system was also developed. First the ITS in the KI group, later work was carried out on a successor to the CTSS system called Multics . In 1970 the AI ​​group separated and the AI ​​Lab at MIT was created, under the direction of Minsky, who took many scientists with him from the MAC project, which merged into the Computer Science Lab at MIT. The AI ​​Lab had already become a center of AI research with worldwide attention at the end of the 1960s. In 1972 Minsky handed over the management of the AI ​​Lab to Patrick Winston . Minsky was also a member of MIT's Media Lab in the 1980s. Later he was there Toshiba Professor of Media Arts and Sciences and Professor of Electrical Engineering and Computer Science (Computer Science).

Minsky has received many awards during his research career. He was a member of the American National Academy of Engineering , the American Academy of Arts and Sciences (since 1968) and the National Academy of Sciences (since 1973). In 1969 he won the Turing Prize , the Japan Prize in 1990 , the Benjamin Franklin Medal in 2001 , the BBVA Foundation Frontiers of Knowledge Award in 2013 and the Dan David Prize in 2014 . He received the RW Wood Prize for his contributions to optics .

Minsky's critics doubt the seriousness of many of his predictions. He predicted that it would soon be possible "to program emotions into a machine". In 1970, he declared that in three to eight years there would be machines with the average intelligence of a person that would read Shakespeare and service cars.

Minsky recently regretted the current development in AI research, as statistical learning methods are increasingly being used here instead of working towards comprehensive modeling of cognitive agents.

Minsky was a member of the Committee for Skeptical Inquiry .

The author Bas Kast , who attended Minsky's Society of Mind seminar as a student , writes of Minsky that he was “a dazzling figure, and hardly anyone who meets him can avoid the impression that it is him a genius acts ” .

His doctoral students include Manuel Blum , Daniel Bobrow , Carl Hewitt , Scott Fahlman , Danny Hillis , Joel Moses , Gerald Jay Sussman , Terry Winograd , Berthold Horn , James Slagle , Patrick Winston and Eugene Charniak .

On January 24, 2016, he died of an intracerebral hemorrhage . He was 88 years old.

The Society of Mind

In his book The Society of Mind (German title: Mentopolis ), Minsky put forward the thesis in 1986 that intelligence consists of an interwoven network of unintelligent agents. Intelligence emerges only through the cooperation of relatively simple agents. Minsky tries to dissuade the reader from the common notion that the human brain is a single, large, monolithic being that thinks about something or is not thinking about something. Instead, a model is sketched in which the brain is made up of innumerable, diverse, but relatively simple agents.

These agents have simple tasks and goals. Only through communication with each other and negotiating the needs of the individual agents with each other do thinking and action arise.

Individual agents can again consist of even smaller agents, which in turn communicate and negotiate with one another. The little agents specialize in certain tasks, for example the eye for sight or a part of the brain to ensure enough sleep. Conflicts within an agent weaken that agent, giving other agents the upper hand.

In this model, learning consists in improving communication between the agents. Personal characteristics of people result from the different weightings of the agents. The storage of memories is made possible by the creation of K-Lines. These K-Lines are a kind of list that contains all agents who were involved in an activity.


  • Neural Nets and the Brain Model Problem , Dissertation, Princeton University, 1954.
  • Computation: Finite and Infinite Machines , Prentice-Hall, 1967.
  • Semantic Information Processing , MIT Press, 1968.
  • Perceptrons , with Seymour Papert, MIT Press, 1969.
  • Artificial Intelligence , with Seymour Papert, Univ. of Oregon Press, 1972.
  • Robotics , Doubleday, 1986.
  • The Society of Mind , Simon and Schuster, 1987.
  • The Turing Option , with Harry Harrison , Warner Books, New York 1992, ISBN 0-446515655 .
  • The Emotion Machine , Simon & Schuster, New York 2006, ISBN 978-0-743276641 .


Web links

Commons : Marvin Minsky  - collection of pictures, videos and audio files

Individual evidence

  1. ^ A b Glenn Rifkin: Marvin Minsky, Pioneer in Artificial Intelligence, Dies at 88. In: The New York Times. January 25, 2016. Retrieved January 25, 2016 .
  2. ^ Biographical Encyclopaedia of American Jews, 1935 , p. 385
  3. ,!5270090/ , / marvin-minsky-obituary
  4. Current Biography Yearbook - 1989 - page 399
  5. Bas Kast: Revolution in the head, the future of the brain. Berliner Taschenbuchverlag, Berlin 2003, ISBN 978-3-442-76150-0 (= instructions for use for the 21st century ).