David Ferrucci

Principal investigator for the Watson/DeepQA Jeopardy! project

• Born: August 11, 1961
• Place of Birth: New York, New York

Primary Company/Organization: IBM

Introduction

Despite his early decision to become a physician, David Ferrucci's lot in life was destined to deal only indirectly with medicine. Instead, he became an expert in artificial intelligence, particularly in programming computers to understand natural human language. He led the team that designed Watson, the IBM supercomputer that managed to win against two top human competitors over three days of the television game show Jeopardy!in February 2011. Watson provided a decisive victory for Ferrucci and his fellow programmers and for IBM, because it demonstrated that its ability to analyze data had implications in a variety of fields that went far beyond entertaining television audiences. Through Watson, Ferrucci was able to show that computers have the potential for improving—and possibly saving—the lives of humans as well as for improving proficiency and performance in academia and business.

Early Life

David A. Ferrucci was born in the Bronx in New York on August 11, 1961. While few details of his personal life are known, it is public knowledge that he had planned to be a physician. However, at the age of seventeen, while taking a math course at New York's Iona College, he began working on a project in computer programming and became fascinated by the possibilities inherent in programming.

Ferrucci received a bachelor of science degree from Manhattan College in 1983, majoring in biology and minoring in computer science. He spent his spare time writing software code. Finally deciding that he preferred computers to medicine, he graduated two years later with a master's degree in computer science from Rensselaer Polytechnic Institute in Troy, New York. That year, Ferrucci began working as a research/software engineer at IBM in the Expert Systems in Manufacturing division. Two years later, he became the lead software engineer at his alma mater, Rensselaer, in the field of object-oriented computer-aided design.

By 1991, Ferrucci had returned to IBM as a researcher and predoctoral intern, assigned to work in the Automated Configuration Systems division. Specializing in knowledge representation and reasoning, he completed his Ph.D. in computer science at Rensselaer in 1994. The following year, he again returned to IBM, where he was a member of a research team conducting research on expert configuration systems, natural language processing, and architectures for natural language engineering.

Life's Work

As IBM's chief architect for unstructured information management applications (UIMA), Ferrucci works with interpreting natural language such as that of text, speech, videos, and images into language that can be understood by computers. UIMA has become the accepted OASIS standard and serves as an Apache open source standard. UIMA, which employs Java and C++ frameworks, is widely used in the business field as well as in academia. Ferrucci's work on UIMA provided the foundation for what is likely to prove to be Ferrucci's defining project, the Watson/DeepQA Jeopardy! project, for which he serves as principal investigator. The project, which began in 2006 and continued for four years at a cost of some $30 million, gave Ferrucci and his team the ability to program Watson to communicate with humans in natural language.

Ironically, the idea for creating Watson was the result of a casual discussion about artificial conversation between Charles Lickel, an IBM executive, and a group of friends during a restaurant meal. The group witnessed fellow patrons becoming enthralled while watching Ken Jennings, the long-running Jeopardy! champion, on the restaurant's television. At the time, Ferrucci was already involved in the field of Deep Questioning Answering (Deep Q/A) at IBM, and he agreed to work on turning the concept of a computer that could win at Jeopardy! into reality. IBM was already working with a number of universities on projects designed to use DeepQA to test the ability of computers to interpret normal human communication methods. IBM had also been involved in a less ambitious project a few years earlier when it created the supercomputer Deep Blue to challenge world chess champion Garry Kasparov in 1996. The computer lost the original match, but a rematch in 1997 between Kasparov and a much-improved model of Deep Blue resulted in a victory for artificial intelligence.

The computer that came to be known popularly as Watson was named after the first president of IBM, Thomas J. Watson. The popular television quiz show Jeopardy!, which debuted in 1964, appeared to be the ideal venue for testing Watson's communicative and interpretive abilities and comparing them with those of human beings, because the show requires rapid access to information, analysis of possible responses at lightning speed, and the ability to remember to place respond in the required question format. Headed by Ferrucci, a team of approximately thirty researchers began programming Watson with Questioning Answering software to prepare Watson for the upcoming challenge. The team was composed of experts in natural language processing, software architecture, information retrieval, machine learning, and knowledge representation and reasoning. Initially, the process of programming Watson as a future Jeopardy! competitor was slow. It took approximately two hours for the computer to arrive at an answer, and it achieved only 15 percent accuracy. Thousands of language analysis algorithms were used to improve Watson's accuracy, and a correct response was based on agreement among at least half of those algorithms.

The result of the efforts of Ferrucci and his team was that Watson was eventually programmed to consider all possible answers to questions likely to appear on the game show and quickly determine the best answer, ultimately achieving 95 percent accuracy and answering questions, on average, within less than three seconds. In practice, Watson was not always able to understand acronyms and abbreviations and had some slight difficulty with concepts from popular culture. Also, Watson is a self-contained unit; it was not connected to the Internet during the game. Thus, its answers were derived only from information available in its own data banks, which consisted of 200 million pages of data that had been programmed into the machine. Unlike Deep Blue, which was dismantled after its rematch with Kasparov, IBM intended to continue to develop Watson's potential.

Even while applauding the success of the Ferrucci team and IBM's work with artificial intelligence, some critics were quick to point out that the Watson team had failed to produce a computer that could think for itself. Ferrucci acknowledged that Watson was unable to think for itself, noting that it was simply a machine that had been programmed to complete particular tasks. He did not argue with the common conclusion that no computer had ever been invented that could match the flexibility of the human brain. However, he saw Watson's Jeopardy! win as a victory for the IBM team and stated that the feat could be considered a “pinnacle” in the field of artificial intelligence.

The implications of the Watson Project are far-reaching, with major potential for success in a variety of fields that include health care, finance, education, law, publishing, and customer service. The implications for the health care field may be the most significant, because Watson's abilities are already allowing medical professionals to identify the best health treatments for particular individuals, using existing medical knowledge and case histories of patients suffering from stipulated conditions and drawing on all available genomic and molecular data.

In September 2011, Watson was put to commercial use when WellPoint Health Insurance began using the supercomputer to determine the proper treatment for its client base of 34.2 million. WellPoint is basing treatment on analysis of charts, medical records, and treatment histories in conjunction with information derived from the database programmed into Watson. In the spring of 2012, IBM and Memorial Sloan-Kettering announced that they had formed a partnership involving Watson for facilitating cancer research. In March 2012, IBM announced that it had created the Watson Healthcare Advisory Board, a nine-member board composed of outstanding physicians, in order to use Watson to work with the medical community in the fields of primary patient care, oncology, biomedical informatics, and medical innovation.

Ferrucci left IBM in 2012 to join the investment firm Bridgewater Associates. In 2015 he became the founder, CEO, and chief scientist of the artificial intelligence company Elemental Cognition. In 2023, Ferrucci announced that he had raised almost $60 million in start-up money for Elemental Cognition.

Personal Life

In 2011, IBM honored Ferrucci by designating him an IBM Fellow, the highest technical distinction awarded by the company. Only 238 other individuals have been awarded that honor since 1963, when it was created, and only 77 of those individuals were working for IBM at the time they were honored.

Bibliography

Arndt, Rachel Z. “Robo Force.” Fast Company 154 (2011): 57–60. Print.

Feldman, Susan. “IBM Watson: From Winning Games to Saving Lives.” 26 Mar. 2012. IDC Link. Web. 2 Sept. 2012.

Ferrucci, David. “One Minute with … David Ferrucci.” Interview by Justin Mullins. New Scientist 209.2799 (2011): n. pag. Print.

Field, Hayden. "The Scientist Behind IBM Watson Has Raised $60 Million for His AI Startup in New York." CNBC, 17 Aug. 2023, www.cnbc.com/2023/08/17/scientist-behind-ibm-watson-raised-60-million-for-elemental-cognition.html. Accessed 6 Mar. 2024.

Kroeker, Kirk L. “Weighing Watson's Impact.” Communications of the ACM 54.7 (2011): 13–15. Print.

Livingstone, Paul. “Far More than Trivial.” R&D Magazine 53.7 (2011): 12–14. Print.

Lohr, Steve. "David Ferrucci: Life after Watson." The New York Times, 6 May 2013, bits.blogs.nytimes.com/2013/05/06/david-ferrucci-life-after-watson/. Accessed 28 Oct. 2019.

Mack, Tim. “Watson: Pathfinder for Global Business.” World Future Review 3.2 (2011): 26–36. Print.

Pieraccini, Robert. The Voice in the Machine: Building Computers That Understand Speech. Cambridge: MIT, 2012. Print.