- Born
- 7 Sep 1941, Kennewick, Washington
- Died
- 3 Jun 2019, Kent, Washington
- Parents
- Roger Gilkey Records (1910–1956), Ethel Madeline Records (1912)
- Spouse
- Carol Annette Records (1942) m. 11 Aug 1963
- Children
- Andrew Neal Records (1969), Laurie Michelle Records (1973)
- Siblings
- Jane Ruth Records (1933–1933), June Arlene Records (1945), George Howard Records (1954)
Roger Michael Records was born in Kennewick, Washington, and graduated from Kennewick High School in 1959. He earned a BA in mathematics from Whitworth College, where he met his wife, Carol, and later an MS in chemistry from Seattle University.
He taught high school chemistry and science for about twenty years, mostly in Federal Way, Washington. He was named Washington State Science Teacher of the Year and received the American Chemical Society's Outstanding Chemistry Teacher award. By every account, he loved the classroom.
To provide for his family, Roger made a mid-career change to the Boeing Company, where he worked in computing and specialized in artificial intelligence. This was the expert-systems era of AI, well before the machine-learning boom of the 2010s, and Roger worked at the leading edge of it.
His research centered on fault management for commercial flight decks: getting a system to reason about many simultaneous aircraft faults and present the crew with a single coherent response instead of a flood of conflicting alerts. Under NASA contract NAS1-18027 he co-authored two NASA Contractor Reports, 'Flight Deck Engine Advisor' (1992) and 'Spurious Symptom Reduction in Fault Monitoring' (1993). In 1994, with Jai J. Choi, he published 'Spurious Symptom Reduction in Fault Monitoring Using a Neural Network and Knowledge Base Hybrid System' at AAAI-94, one of the field's top artificial intelligence conferences. That same year he was named an inventor on Boeing patent US 5,522,026, a system for consolidating multiple faults into a single electronic checklist. His pairing of neural networks with knowledge-based reasoning was, in hindsight, a bridge between the symbolic AI of his era and the machine learning of ours.
Boeing named him an Associate Technical Fellow in recognition of his research in computing, a distinction held by only a small fraction of the company's technical staff.
Family recollection: at a teachers' conference, Jeff Records's father was seated among a group of Boeing engineers who grew excited that a 'Roger Records' was a listed speaker, only to discover it was a different man. They spoke of Roger M. Records's Boeing work with real admiration. When Jeff's father took their praise to mean Roger was simply a nice fellow, an engineer corrected him: not a nice fellow, a Fellow at Boeing.
Roger's deepest commitment was his faith. He came to faith following high school and served for the rest of his life, teaching Sunday school, serving in church leadership, and acting as a Teaching Leader in Bible Study Fellowship for eleven years. In retirement he collected marbles and Erector sets, and, as his family fondly remembers, never met a banana cream pie he did not like.
From Jeff Records: I work today as an AI architect, and understanding this corner of computer science has always come naturally to me. It means something to learn that a cousin was doing AI work at Boeing at such a high level a generation before the field became what it is now. The parallel runs deeper than the subject. Roger trained and worked as a teacher before financial and job-market realities moved him into technology, where he thrived; I trained as an elementary educator and took a job in tech out of the same kind of constraint. My brother, James Michael Records, is also deeply involved in AI in his work at F5, and he and I have built a few things together. Across the family, the pull toward teaching machines to reason keeps resurfacing.