Raj Reddy: the boy from Katoor who taught computers to listen

From Katoor to Australia

Dabbala Rajagopal Reddy was born on 13 June 1937 in Katoor, Andhra Pradesh. His father was an agricultural landlord and his mother a homemaker. He went to the ZP High School at Sri Kalahasti in Chittoor district and earned a degree in civil engineering from Guindy College of Engineering in Madras, now part of Anna University, in 1958.

He then went to Australia as an exchange student and completed a master's degree in technology at the University of New South Wales in Sydney in 1960. For the next three years he worked for IBM in Australia as an Applied Science Representative. The ACM's biography notes that his first encounter with a computer was an English Electric DEUCE.

A class project that lasted a lifetime

In 1963 Reddy joined Stanford University as a PhD student. In early 1964 he began a class project on speech recognition under John McCarthy, one of the founders of artificial intelligence. He used the Stanford AI Lab's new equipment to turn speech waveforms into numbers a PDP-1 computer could process. He later said he chose that project because he was interested in natural languages and what computers could learn about them.

He finished his PhD in 1966, on speech recognition. It was the first PhD awarded by Stanford's newly formed Department of Computer Science. He stayed on as an assistant professor, working on speech, image processing and face recognition.

Carnegie Mellon and the Turing Award

In 1969 he moved to Carnegie Mellon University in Pittsburgh. There he led the university's part of a large speech understanding project funded by DARPA. His teams built a series of systems: Hearsay I, Hearsay II, Harpy and Dragon. Hearsay I was the first system capable of recognising continuous speech, and the ACM says most of the ideas behind modern commercial speech recognition came from this line of work. The blackboard model, a way of letting many sources of knowledge work together, spread across applied AI.

Over three decades he and his colleagues demonstrated voice control of robots, large-vocabulary connected speech recognition, speaker-independent recognition and unrestricted dictation. He became founding director of CMU's Robotics Institute, served as dean of its School of Computer Science from 1991 to 1999, and later led CMU's West Coast campus.

In 1994 he and Edward Feigenbaum received the ACM Turing Award for pioneering large-scale artificial intelligence systems and showing the practical and commercial value of AI. Indiaspora notes that he was the first person of South Asian origin to receive the award. India gave him the Padma Bhushan in 2001, and Japan's Honda Foundation awarded him the Honda Prize in 2005.

He kept a link with India too. He served as chairman of the International Institute of Information Technology, Hyderabad, and his research interests at CMU include voice computing for people who cannot easily read and write.

The skills and habits that made the difference

The first was following curiosity. Reddy picked speech recognition from a list of possible projects because natural language interested him, and that interest carried him for decades.

The second was building working systems. His fame comes not from theory alone but from programs that actually listened, from Hearsay to Dragon, each one improving on the last.

The third was moving between disciplines. He began in civil engineering, learned computing on the job at IBM, and then joined the young field of AI. Each step built on what came before.

The fourth was teamwork. The ACM credits him and his students together. Much of his impact came through the people he trained and the institutes he led.