Last year we lost Robert Amsler, a valued member of our field, on August 12, 2025. Bob was a computational lexicologist, a computational linguist, and an information scientist. After an MS in Mathematics and CS from the Courant Institute, NYU, he went on to a Ph.D. in Computer Science (1980) at UT-Austin, with Robert Simmons as his advisor. I was fortunate enough to share an office with him at the University of Texas in 1973/1974 and benefitted from his kindness, his intelligence, and his meticulous attention to detail for the next 50 years. Thanks to Bob’s generosity, his other office mate, Mike Smith, colleague Clive Dawson, and I have been able to establish The Robert Amsler and Robert Simmons Endowment for Texas Robotics at the University of Texas, Austin.
https://www.linkedin.com/posts/texasrobotics-longhornsgive-utaustin-shar...
After his PhD in CS at the University of Texas in Austin, he worked at SRI (Stanford Research Institute), Bellcore (Bell Labs) and the MITRE Corporation. He ended his career at the U.S. Department of Energy, where he worked on projects to develop automated declassification software for documents and to design a history knowledge base for DOE’s nuclear weapons program.
Robert was always fascinated by words and word meanings, and this sparked an early interest in vast collections of words, such as those found in dictionaries. He was convinced that the success of getting computers to understand natural language would depend on having a broad, expansive coverage of word meanings and how the meanings changed in different contexts. Anything Robert specialized in became a very deep endeavor motivated by an insatiable scientific curiosity. Data exploration, in particular, was his forte.
His early efforts included work with machine-readable dictionaries. He did considerable research using the Merriam-Webster Pocket Dictionary, which ended up being the subject of his doctoral dissertation. In the early 1970s, he had to obtain the dictionary from the publisher on thousands of feet of punched paper tape, which took dozens of hours to upload into the computer. The results were seminal. Robert was able to perform a wide variety of analyses of English based on the noun and verb taxonomies he extracted from the dictionary that had not been previously possible in an academic setting, helping to motivate the creation of WordNet.
He was equally interested in going in the other direction, investigating how the usage of words in context, found in large bodies of text, could support new word meanings that could be added to dictionaries. This sparked a career in encouraging the release of machine-readable dictionaries, subsequent explorations of the data contained therein, and a long-standing partnership with Don Walker at both SRI and Bellcore. In their studies they anticipated by more than 40 years the success of large language models (LLMs), which today form the backbone of modern generative AI systems such as ChatGPT or Gemini. They process large amounts of text in order to infer implicit semantic relations between words, just as Bob and Don were doing. Robert also invented a new technique in citation analysis that bears his name. His work is mentioned in Wikipedia articles on Machine-Readable dictionaries, Computational lexicology, Bibliographic coupling, and Text mining.
He is survived by neighbors, friends, and colleagues who remember him fondly for his brilliant mind, joyful spirit, and gentle heart.