Jaime Guillermo Carbonell (July 29, 1953 – February 28, 2020) was a computer scientist who made seminal contributions to the development of natural language processing tools and technologies. His research in machine translation resulted in the development of several state-of-the-art language translation and artificial intelligence systems. He was the Allen Newell Professor of Computer Science and head of the Language Technologies Institute at Carnegie Mellon University, where he worked from 1979 until his death.
Carbonell's work spanned multiple areas of artificial intelligence, including machine learning, text mining, and natural language processing. He created the MMR (maximal marginal relevance) technology for text summarization and novelty detection, invented transformational analogy for case-based reasoning, and developed knowledge-based interlingual machine translation. He also co-founded the field of modern machine learning with Ryszard Michalski and Tom Mitchell, organizing the first machine learning conferences and co-editing the first books on the subject.
Early Life and Education
Carbonell was born on July 29, 1953. He completed his undergraduate studies at the Massachusetts Institute of Technology, receiving dual B.S. degrees in Physics and Mathematics in 1975. He then pursued doctoral studies at Yale University under the supervision of Dr. Roger Schank, completing his Ph.D. in computer science in 1979. His dissertation work focused on natural language understanding and case-based reasoning, laying the groundwork for his later innovations.
Career at Carnegie Mellon
Carbonell joined Carnegie Mellon University as an assistant professor of computer science in 1979, moving to Pittsburgh to begin his academic career. He was appointed full professor in 1987, becoming the youngest chaired professor in the School of Computer Science at the time. In 1995, he was awarded the Allen Newell endowed chair, and in 2012 he was named University Professor, a rare distinction recognizing his exceptional contributions.
He was affiliated with several departments at Carnegie Mellon, including the Language Technologies Institute, Computer Science Department, Machine Learning Department, and Computational Biology Department. Carbonell served as the founding director of the Language Technologies Institute, which grew to become the preeminent global institution for language studies, with over $12 million in annual funding and 200 researchers. The institute achieved top honors in machine translation, search engines, speech synthesis, and education, and its model was later adopted by institutions in Germany (DFKI), Japan (Tokyo University), and the United States (Johns Hopkins).
Research Contributions
Carbonell's research interests were broad, spanning artificial intelligence, language technologies, and machine learning. He made major technical contributions in several fields, including the creation of MMR technology for text summarization and informational novelty detection in search engines. This work became foundational for modern information retrieval systems, influencing how search engines rank and present results.
He invented transformational and derivational analogy, generalized methods for case-based reasoning that allow systems to reuse, modify, and compose past successful plans for increasingly complex problems. In machine translation, he developed knowledge-based interlingual approaches and later worked on trainable example-based and statistical methods, including a symmetric optimal phrasal alignment method. His teams achieved several notable firsts: the first scalable high-accuracy interlingual machine translation system in 1991, the first speech-to-speech machine translation system in 1992, and the first large-scale spider and search engine in 1994.
Carbonell also contributed to bioinformatics, developing linked conditional random fields for predicting tertiary and quaternary protein folds. He was instrumental in setting up the Computational Biolinguistics Program, a joint venture between Carnegie Mellon and the University of Pittsburgh that combined language technologies and machine learning to model genomic, proteomic, and glycomic 3D structures. His team achieved the first trainable, large-scale protein-structure topology predictor in 2005.
Machine Learning Pioneer
Carbonell was a co-founder of modern machine learning, working alongside Ryszard Michalski and Tom Mitchell to establish the field as a distinct discipline. Starting in 1980, he co-edited the first three books on machine learning, including the seminal "Machine Learning: An Artificial Intelligence Approach" (1983) and its second volume (1986). He organized the first four machine learning conferences, beginning with the 1981 workshop at Carnegie Mellon, and was a co-founder and editor-in-chief of the Machine Learning Journal.
His work in machine learning included proactive machine learning for multi-source cost-sensitive active learning, seeded version-space learning, and series-anomaly modeling for financial fraud detection and syndromic surveillance. These contributions helped establish machine learning as a fundamental enabling technology in search engines, data mining, and social networking.
Entrepreneurship and Impact
Carbonell's innovations led to several successful start-up companies. He was involved with Carnegie Group, which focused on AI expert systems; Lycos, a web search engine founded by his Ph.D. student Michael Mauldin; Wisdom, which applied financial optimization and machine learning; Carnegie Speech, which developed spoken-language tutoring systems; Dynamix, which specialized in data mining and pattern discovery; and Meaningful Machines, which worked on context-based machine translation. These ventures demonstrated his ability to translate academic research into practical applications with commercial impact.
Awards and Honors
Carbonell received numerous awards throughout his career. He was elected a fellow of the Association for the Advancement of Artificial Intelligence (AAAI) in 1991. He received the Okawa Prize in 2015 for his contributions to information and telecommunications research. In 1997, he won a best paper award for "Translingual Search" with Yang at the International Joint Conference on AI. Earlier in his career, he received the Herbert Simon teaching award in 1986, the Sperry Fellowship for excellence in AI research in 1986, and the Computer Science teaching award at Carnegie Mellon in 1987. He also served as president of the ACM Special Interest Group on Artificial Intelligence (SIGART) from 1983 to 1985, receiving a "Recognition of Service" award from the ACM. In 1990, he provided congressional testimony on machine translation, reflecting his influence on national policy discussions.
Legacy
Carbonell wrote more than 300 technical papers and gave over 500 invited or refereed presentations during his career. He died following a long illness on February 28, 2020. His legacy continues through the Language Technologies Institute at Carnegie Mellon, which remains a leading center for natural language processing research. His contributions to machine translation, text summarization, and machine learning have had a lasting impact on the field, influencing subsequent developments in deep learning and large language models. In 2023, Mona Talat Diab became the director of the Language Technologies Institute, continuing the tradition of leadership that Carbonell established.