In Memoriam

Remembering Fred Thompson (1922-2014)

Written by Philip M. Neches BS73ES, PhD83CS, additional material from Peter Szolovits BS70Ph, PhD75ES

Fred Thompson was my undergraduate advisor when I switched my major to Engineering as a junior, and was, with Carver Mead, my PhD thesis advisor. He passed away May 27, 2014, in his 92nd year.

Fred was unusual, even for a Caltech professor. Short and thin, he could resemble a greying Puck when clean-shaven or the Arthurian wizard Merlin in full beard. The Merlin effect worked particularly well when he donned academic regalia for Commencement.

To Fred, the unknown variable was not “x”. Rather, it was “bunny rabbit”. “Why?” his students would ask. Because it was as good as any other name for a variable.

Born Frederick Burtis Thompson on July 26, 1922, Fred served in the Army and worked at Douglas Aircraft during World War II. He earned his bachelors (1946) and masters (1947) degrees at UCLA and his PhD (1951) at Berkeley, all in mathematics. He joined RAND Corporation after finishing his studies under logician Alfred Tarski. At RAND, Fred is credited with inventing discrete event simulation.

While at RAND, Fred worked on a study of what would happen to American cities in the aftermath of a thermonuclear attack. The study sought to provide recommendations on how to prepare emergency services for such an eventuality. When the team developed their answer, Fred was in tears: the destruction would be so devastating that no services would survive, even if a few people did. Fred confronted the unthinkable, and came away a changed person. But this kind of hard-headed analysis eventually led policy makers to a simple conclusion: the only way to win a nuclear war is to never have one.

Mathematical models derived from the bomb models Fred helped develop have been used in recent years to help optimize placement of medical specialists to maximize care delivery and to optimize delivery of emergency medical supplies in the aftermath of the Haiti earthquake. The models treat the doctors and supplies as the bombs and calculate the number of people affected. Life has its ironies, and Fred would be the first to appreciate them.

Fred moved to GE Research in 1957. While at GE, he began what would become his long-term interest in natural language as the ideal way for people to interact with computers. Fred’s first effort to teach English to a computer was a system called DEACON, developed in the early 1960s.

Fred joined the Caltech faculty in 1965. He advised the computer club as a canny way to recruit a small but dedicated cadre of students to work with him.

He began a collaboration with Bozena Henisz-Dostert, an accomplished linguist and wife of machine translation researcher Leon Dostert. Leon Dostert died in 1971, aged 67. Bozena was a widow at 36. In 1969, Fred and Bozena began a lifelong collaboration which was personal as well as professional; their wedding was the second marriage for each. Together, they pioneered the application of deep linguistic theory to the problem of natural language processing by computer. Working with a generation of Caltech’s best and brightest undergraduate and graduate students, they created REL (Rapidly Extensible Language) in the 1970s and POL (Problem Oriented Language) in the 1980s. The data representation in REL (object – attribute – value – with time spans and open/closed interval markets) foreshadowed today’s semantic web representations.

Bozena was a tall, aristocratic Polish blonde. Fred and Bozena never had children. They raised and showed Borzois: immense Russian wolfhounds. Each dog probably weighed as much as Fred. Bozena walking a pair of them on the streets of Pasadena around the Huntington hotel was a striking sight indeed. Caltech never granted Bozena tenure, despite her accomplishments and reputation as a linguist. Caltech’s treatment of tenure cases for female faculty members in the 1970s was not the Institute’s finest hour. Bozena passed away in 2002.

Fred taught both the most theoretical and the most practical computer science courses at the Institute, long before Caltech had a formal computer science department. In his theory class, students proved the equivalence of a computable function to a recursive language to a Turing machine. In his data analysis class, students got their first appreciation of the growing power of the computer to handle volumes of data in novel and interesting ways. Careers were launched simply by taking Fred’s IS 142 course.

One demonstration of REL went so well that the sponsor thought the whole thing was canned. Fred found out only when he inquired as to why his funding was not renewed. From then on, each demo started with a part that was rehearsed, so it would work, and then an open-ended session, practically guaranteed to run into a bug or an unimplemented feature. That way, the reviewer of the research software knew it was the real thing. Characteristic of Fred, he shared the lesson openly with his students.

As a linguist, mathematician, and philosopher, Fred believed that humans have an internal model of their world. He believed that his model was recursive, therefore computable. People use their internal model to classify and make sense of the data they receive from what they see, hear, and feel. This includes what they read or hear from other people. Without a framework, there is no way to make sense of the input. This theory underpinned his work on computer understanding of human language.

Fred postulated that deep learning, like understanding a new theory of science or a new perspective on life, constituted a change in the internal model, going from one recursive system to another. Further, he postulated that the process of that change was not recursive, that is to say, not computable. Put another way, Fred believed that deep human learning and creativity could never be simulated by a computer. Fred worked in the background for decades to prove this assertion, one that would stand with Gődel’s incompleteness theorem in importance. I’m not sure that he ever did prove his grand theorem.

Fred was one of the three founding professors of the Computer Science department at Caltech in 1976. While Carver Mead and Ivan Sutherland thought about how VLSI would impact how information systems are built, Fred kept the original cadre of seven graduate students focused on what computers were for.

Today, we talk about how computation has so changed how science and engineering are done. We rank computation as the third leg of the stool of the scientific method, along with theory and experiment. The popular term for the field has evolved from Management Information Systems to Decision Support Systems to Business Intelligence Systems to Data Mining to simply Big Data. Fred and his students pioneered the arena starting more than 50 years ahead of the pack.

Fred is survived by his third wife, Carmen Edmond-Thompson of Altadena, CA. They married in 2002, enjoying travel and dancing. They discussed moving to England, where Carmen, a native of Jamaica, grew up. Fred is also survived by two children from his first marriage, Dr. Mary Ann Thompson, Director of Hematology at Vanderbuilt, and Scott Thompson of Chicago. He is also survived by four grandchildren.

Fred was a gentle soul and a gentleman, with an abiding passion for his work and for people he cared about. His warmth and humility complemented a wide-ranging and towering intellect.

Fred viewed his students as his intellectual children. He sought to equip us with a keen sense of inquiry, a taste for large and important problems, and a deep understanding of human realities. Nothing made him happier or prouder than to learn of our success.

Allen E. Puckett (1919-2014)

Allen E. Puckett (PhD ’49), the engineer who helped father the delta- winged airplane, the guided missile, and the communications satellite, passed away at his home in Pacific Palisades, California, on March 31, 2014.

He earned his bachelor’s and master’s degrees in engineering at Harvard (in 1939 and 1941, respectively) before coming to Caltech to pursue his doctorate in aeronautics under Theodore von Kármán, the leading aerodynamicist of the era. Puckett’s PhD thesis, “Supersonic Wave Drag on Thin Airfoils,” laid the foundation for designing the triangular-shaped delta wings found on such diverse aircraft as supersonic fighter jets, the SR-71 Blackbird spy plane, and the Space Shuttle orbiter.

Among other honors, Puckett won the Lawrence Sperry Award of the Institute of Aeronautical Sciences (now the American Institute of Aeronautics and Astronautics) in 1948. He was named a Caltech Distinguished Alumnus in 1970, the California Manufacturer of the Year in 1980, a Chevalier of the French Legion of Honor in 1984, and was awarded the National Medal of Technology by President Reagan in 1985. At Caltech, Puckett endowed a chair in the Division of Engineering and Applied Science. Robert McEliece is the Allen E. Puckett Professor and Professor of Electrical Engineering, Emeritus; Pietro Perona is the Allen E. Puckett Professor of Electrical Engineering. Caltech’s Guggenheim Aeronautical Laboratory, the building where Puckett spent his time on campus as a grad student, was extensively renovated in 2008. The west end of the third floor now houses the Allen Puckett Laboratory of Computational Fluid Mechanics. Puckett is survived by Marilyn Puckett, his wife of 50 years, five children, six grandchildren, and 14 great-grandchildren.

Paul H. Patterson (1943-2014)

Paul H. Patterson, the Anne P. and Benjamin F. Biaggini Professor of Biological Sciences, Emeritus, at Caltech, and a neuroscientist and developmental biologist who created novel behavioral models of schizophrenia and autism in mice, died on June 25, 2014.

Born in Chicago in 1943, Patterson stayed in the Midwest for college, graduating with a bachelor’s degree from Grinnell College in Iowa in 1965. From there, he moved east for graduate school at Johns Hopkins University, earning his doctorate under advisor William Lennarz in 1970. In 1983, after more than a decade as a faculty member at Harvard Medical School, Patterson joined the faculty at Caltech.

His research focused on inter- actions between the nervous and immune systems—a connection that was not universally acknowledged in the early days of neuroscience. “Professor Patterson was a pioneer and iconoclast who was not afraid to work outside the scientific mainstream, and who consequently made a number of important and seminal contributions that opened up entire fields of research,” says David Anderson, Seymour Benzer Professor of Biology at Caltech and Patterson’s longtime colleague.

A mouse model he developed has been used to study the environmental factors that influence the symptoms of human neurodevelopmental disorders and has increased awareness of the importance of those influences. Recently, the model informed a possible new therapy to treat schizophrenia- associated hallucinations. In another recent study, Patterson and colleagues demonstrated that the gut microbiome, the diverse collection of bacteria that reside in the intestine, regulates behaviors in a mouse model of autism.

Patterson also contributed to the understanding and treatment of Huntington’s disease, a devastating hereditary neurological disorder, and he was instrumental in developing the Institute’s MD/PhD joint degree program—a collaboration that allows graduate students to combine their Caltech research experience with medical education at UCLA or USC. He is survived by his wife, Carolyn, and his son, Paul.

Frank E. Marble (1918-2014)

Frank E. Marble, the Richard L. and Dorothy M. Hayman Professor of Mechanical Engineering and Professor of Jet Propulsion, Emeritus, at Caltech, passed away on August 11, 2014.

Marble received his bachelor of science degree in 1940 and his master’s degree in 1942, both from the Case Institute of Technology. He then came to Caltech and earned an engineer’s degree in 1947 and a PhD in 1948, with Professor Theodore von Kármán as his advisor. He was hired at Caltech in 1948 as an instructor in aeronautics, became assistant professor of jet propulsion and mechanical engineering in 1949, associate professor in 1953, professor in 1957, and was named Hayman Professor of Mechanical Engineering and Professor of Jet Propulsion in 1980. He retired in 1989.

Marble made major contributions to aerodynamics, combustion, and propulsion, specifically the research and development of gas turbines and rockets. He also was responsible for the training of several generations of scientists in the field of aeronautics.

A member of both the National Academy of Engineering and the National Academy of Sciences, Marble received many honors, including the 1999 Daniel Guggenheim Medal, awarded by the American Institute of Aeronautics and Astronautics (AIAA), and the AIAA Combustion Award. Marble was predeceased by Ora Lee, his wife of seven decades.