Georgetown experiment

The Georgetown Experiment - a demonstration of machine translation capabilities, held January 7, 1954 in New York , at the headquarters of IBM Corporation. Prepared by Georgetown University in conjunction with IBM. During the experiment, a fully automatic translation of more than 60 sentences from Russian into English was demonstrated. The presentation had a positive impact on the development of machine translation in the next 12 years.

The experiment was conceived and prepared in order to attract public and government attention. Paradoxically, it was based on a fairly simple system : it was based on only 6 grammatical rules , and the dictionary included 250 entries. The system was specialized: organic chemistry was chosen as the subject area for translation, and a number of general proposals were also added. The program was run on the IBM 701 mainframe . Proposals like “Processing improves the quality of oil”, “Commander receives information by telegraph” were entered into the computer on a punch card , and the machine would output their translation printed by the translator .

The experiment was led by a professor at Georgetown University, Leon Doster, and the head of applied research at IBM, Cuthbert Heard. The immediate conduct of the experiment was the responsibility of associate professor at Georgetown University, Paul Garvin, for the linguistic part, and IBM's employee Peter Sheridan, for the computer part.

The demonstration was widely reported in the media and perceived as a success. It influenced the decision of some governments , primarily the United States , to invest in computational linguistics .

In the same year, the first experiment on machine translation was carried out in the USSR , at the Institute of Precise Mechanics and Computer Engineering of the USSR Academy of Sciences , on a BESM computer. Isabella Belskaya led the research, initiated by the director of the Institute Dmitry Panov. In parallel with the Belskaya group, a group of scientists from the Applied Mathematics Department of the Mathematical Institute of the Academy of Sciences of the USSR named after V.A. Steklov under the direction of Olga Kulagina and Alexey Lyapunov .

The organizers of the Georgetown experiment assured that within 3-5 years the problem of machine translation would be solved. However, in reality, everything turned out to be more complicated, and in 1966, based on the report , which concluded that more than 10 years of research did not yield a complete result, funding was significantly curtailed.