IBM Research

What it is

IBM Research is the research division of IBM. Its headquarters is the Thomas J. Watson Research Center in Yorktown Heights, New York. In 2021 it had twelve laboratories on six continents.^[1] Its researchers have been credited with six Nobel Prizes and seven Turing Awards.^[1]

Origins

The division's roots go back to 1945, when the Watson Scientific Computing Laboratory opened at Columbia University. It was IBM's first laboratory devoted to pure science.^[1] Its link to the early history of AI goes back to 1955. Nathaniel Rochester of IBM Corporation was one of the four signatories of the 1955 Dartmouth proposal on artificial intelligence.^[4] IBM opened its first European research laboratory, in Switzerland near Zurich, in 1956.^[1]

Known for

The best-known AI project from IBM Research is Deep Blue. Feng-hsiung Hsu and Murray Campbell had built the chess machine Deep Thought at Carnegie Mellon University. They joined IBM Research in 1989, where the project was renamed Deep Blue.^[2] In May 1997 Deep Blue beat world chess champion Garry Kasparov 3.5–2.5 in a six-game rematch in New York. The machine evaluated 200 million positions per second.^[2]

Chess was a contest of search and evaluation, and Deep Blue's speed was the central fact of its design. Its 1997 win was a public demonstration that a computer could beat the best human at a game long seen as a test of reasoning.

Watson was the second major IBM project to reach the public. In February 2011 IBM's Watson question-answering system won a televised Jeopardy! contest against champions Ken Jennings and Brad Rutter. It finished with $77,147, against their $24,000 and $21,600.^[3] Watson was built at the Thomas J. Watson Research Center by a team of more than two dozen, led by principal investigator David Ferrucci, who pitched the project in 2006. It used IBM's DeepQA software on 90 servers with 2,880 processor cores.^[3]

In 2017 IBM invested $240 million to create the MIT–IBM Watson AI Lab. The lab funds about 50 projects a year in areas including computer vision, natural language processing and reinforcement learning.^[1]

Place in AI history

IBM Research's record shows how AI moved from laboratory work into public view. Deep Blue showed that a computer could beat the world chess champion in a match. Watson showed that a machine could answer quiz clues written in natural language well enough to beat the best human players. Both projects depended on large-scale engineering, not only on new theory. The MIT–IBM Watson AI Lab, set up in 2017, continued that pattern by pairing industrial research with a university partner.