In Memoriam: Charles Rosen, Norman Nielsen, and Saul Amarel

AI Magazine Volume 24 Number 1 (2003) (© AAAI) In Memoriam In Memoriam Charles Rosen, Norman Nielsen, and Saul Amarel Peter E. Hart and Nils J. Nilsson, Ray Perrault, and Tom Mitchell and Casimir A. Kulikowski n the span of a few months, the AI community lost four important figures. The fall of 2002 marked the passing of Ray Reiter, for whom Ia memorial article by Jack Minker appears in this issue. As the issue was going to press, AI lost Saul Amarel, Norm Nielsen, and Charles Rosen. This section of AI Magazine commemorates these friends, lead- ers, and AI pioneers. We thank Tom Mitchell and Casimir Kulikowski for their memorial to Saul Amarel, Ray Perrault for his remembrance of Norm Nielsen, and Peter Hart and Nils Nilsson for their tribute to Charles Rosen. The AI community mourns our lost colleagues and gratefully remembers their contributions, which meant so much to so many and to the advancement of artificial intelligence as a whole. – David Leake 6 AI MAGAZINE Copyright © 2003, American Association for Artificial Intelligence. All rights reserved. 0738-4602-2003 / $2.00 In Memoriam Charles Abraham Rosen, December 7, 1917 – December 8, 2002 SPRING 2003 7 In Memoriam Charles Abraham Rosen— (EBL). These techniques are still foundational to- day, almost 40 years later. Scientist and Visionary: An example of Charlie’s sense of humor is dis- 1917–2002 played in the 1972 SRI film “SHAKEY, the Robot.” Dressed in black cape as a gremlin, Charlie attempts harlie Rosen combined enormous creativity to frustrate SHAKEY by blocking its path to a goal. and a spirit of adventure with an equally (SHAKEY’s planning and execution-monitoring soft- enormous sense of responsibility at all levels. C ware was able, of course, to overcome such unantic- The foundation of Charlie’s creativity was his ipated interference.) broad knowledge. His intellectual curiosity resulted Charlie wanted to do more with robots than just in a head crammed with facts and figures from many use them as a platform for developing AI tech- disciplines. He knew a lot about applied physics, niques. He wanted to develop robot systems that about what used to be called radio engineering, could actually be applied to important problems in about manufacturing engineering, plant physiology, industry. In the mid-1970s, he formed a Robotics neurobiology, adaptive learning algorithms and de- Group within SRI’s AI Center and began work on vices, machine vision, industrial robotics, national several projects aimed at moving robots out of the politics, international affairs, and art. His special cre- laboratory. ative genius was his ability to associate knowledge One of the important inventions of the Robotics from these different disciplines and to recognize the Group at that time was a very robust machine vi- possibilities of combining apparently unrelated ideas sion system capable of recognizing industrial parts in new and interesting ways. as they came down conveyor belts. Ultimately, in- Born in Canada, Charlie received a B.S. degree in terest in this system led Charlie and others to found electrical engineering at Cooper Union followed by Machine Intelligence Corporation, which commer- a master’s at McGill University in Montreal and a cialized this product. Ph.D. at Syracuse University. He came to SRI Inter- Charlie’s sense of responsibility had no bound- national (then called the Stanford Research Insti- tute) in the late 1950s after working with solid-state aries. He felt personally responsible for everyone he devices at the General Electric Company in Syra- worked with, taking special care to mentor younger cuse, New York. While at G.E., he coauthored an in- colleagues. He felt a great sense of business respon- fluential textbook on transistors. sibility to the funders and investors who made his Charlie’s first involvement with artificial intelli- work possible and was always straightforward in his gence was his work on a series of neural network depiction of risks and rewards. And he had an un- machines at SRI during the early 1960s. The last of limited sense of social responsibility, donating, for example, his own time to goals like “appropriate these, MINOS III, achieved impressive results—includ- ing classification of symbols on army maps and ac- technology” for the developing world, which he curate recognition of hand-printed characters on did under the auspices of the National Academy of Engineering. FORTRAN coding sheets. In 1965, Charlie proposed creating the world’s Charlie sometimes observed that there were two kinds of people in the world, “givers” and “takers.” first intelligent mobile robot, and SHAKEY was born. No one ever doubted what kind of person Charles The genius behind SHAKEY was Charlie’s vision of a project that would combine in one experimental A. Rosen was. His passing is a great loss. We shall test-bed advances from every subfield of artificial not soon see his kind again, but his influence and intelligence. That indeed was exactly what hap- example will be with us for a long time to come. We pened. SHAKEY combined machine vision, auto- are proud and privileged to have been able to work matic planning and execution monitoring, naviga- under him. tion, speech and natural language understanding, —Peter E. Hart and Nils J. Nilsson and robot software architecture in ways that were initially unforeseen and had never been done be- Norman Russell Nielsen: 1941–2002 fore. The SHAKEY project, funded by DARPA for several years, generated several very important AI tech- orman Nielsen, the secretary-treasurer of nologies. Among them were the introduction of re- the American Association for Artificial gion finding and the invention of the so-called NIntelligence (AAAI) since 1992, died at his “modified” Hough transform in computer vision, home on 25 December 25 2002. the development of three-level robot software ar- Since 1975, Norm was an information technol- chitectures, the A* heuristic search algorithm ogy consultant for SRI International and its sub- (which among many other applications is univer- sidiaries, where he led projects in applied AI, sally used for computing driving directions), the speech recognition, network architecture, and in- STRIPS automatic planning system, and what has formation security. Because of his deep under- come to be called “explanation-based learning” standing of their real needs, his technical insight, 8 AI MAGAZINE In Memoriam Norman Russell Nielsen, September 8, 1941 – December 25, 2002 SPRING 2003 9 In Memoriam and his integrity, he developed long-lasting rela- determination were legendary in the communi- tionships with clients around the world. He led ty—as an enthusiastic leader, visionary researcher, the effort to create an expert system to assist in the and pioneering contributor to advanced computing selection of fasteners used in automobiles at Gen- and AI methodologies, applied to both scientific in- eral Motors (GM). Not only was the expert system quiry and engineering practice. application successful, but by capturing the Saul Amarel developed seminal concepts and knowledge of how fasteners should be selected, wrote foundational papers on knowledge represen- GM was able to improve its design practices and tation for human and automated problem solving. reduce the number of fasteners used in cars from An early example is “On the Automatic Foundation 5000 to 500, resulting in substantial savings. In a of a Computer Program Which Represents a Theo- project for the Electric Power Research Center, ry” (Self- Organizing Systems, eds. Yovits et al., Spar- Norm pioneered the use of expert systems for the tan, 1962), written while he headed the RCA control of utility power generators, improving the Sarnoff Labs Computer Theory Group. He attended monitoring, control, and safety of power-generat- the Allerton House Conference organized by Von ing plants beyond what was practical with just hu- Foerster of the University of Illinois in that year, man operators. which brought together a number of early AI re- The valedictorian of his class at Pomona Col- searchers inspired by the work of Warren McCul- lege, Norm received an M.B.A. and Ph.D. in oper- logh in his Embodiments of Mind, and later in the ations and systems analysis from Stanford Univer- decade, Amarel became a member of the board of sity, served as assistant director of the Stanford directors of the American Cybernetics Society Computation Center, and taught at the Stanford (ACS), which emphasized the cybernetic origins Graduate School of Business before joining SRI. He and strand within AI. In 1966, Amarel spent a sab- was the author of two books on applied AI and batical at Carnegie Mellon University, where he more than 75 articles. joined Allen Newell’s seminar on shifts in knowl- As secretary-treasurer of AAAI, he chaired the Fi- edge representation. As the result of this seminar, nance Committee, which oversees annual budgets Amarel published “An Approach to Heuristic Prob- and the association’s investments, and was a mem- lem Solving and Theorem Proving in the Proposi- ber of the AAAI Executive Committee. He was a tional Calculus” (Systems and Computer Sciences, eds. valuable source of advice on policy, as well as fi- Hart and Takasu, University of Toronto, 1967), nances, and his quiet words of common sense were which showed how an appropriate representational always respected and appreciated. choice can vastly simplify the search for problem He leaves his wife of 39 years, Jeannette, and a solutions. Amarel’s research evolved at this time daughter, Joanne. from networks, communication, and simulation to- Norm was a seasoned traveler (he visited Japan ward more general AI problem- solving approaches. over 60 times), a devoted outdoorsman, a passion- His sojourn at Carnegie Mellon proved fortuitous, ate supporter of Stanford athletics, and a lifelong leading to lifelong friendships and professional as- lover of trains.

Load more