Speech Pioneer Fred Jelinek Dies at 77

Article Featured Image

Frederick Jelinek, a Johns Hopkins University faculty member whose research laid the foundation for modern speech recognition and text translation technology, died Sept. 14 while working at the university’s Homewood campus in Baltimore. He was 77.

During 21 years at IBM Research and nearly two decades at Johns Hopkins, Jelinek pioneered the statistical methods that enable modern computers to understand,’ transcribe, and translate written and spoken language. In recognition of this work, he was inducted in 2006 into the National Academy of Engineering.

“He envisioned applying the mathematics of probability to the problem of processing speech and language,” said Sanjeev Khudanpur, a Johns Hopkins associate professor of electrical and computer engineering who worked with Jelinek. “This revolutionized the field. Fifty years ago no one thought that was possible. Today, it’s the dominant paradigm.”

Khudanpur added, “Initially, Fred’s ideas generated some hostile reviews, owing to philosophical differences, but ultimately his approach prevailed and became mainstream. Over the past 10 years or so, he received lifetime achievement awards, one after another, from diverse professional societies.”

At Johns Hopkins, which he joined in 1993, Jelinek was the Julian S. Smith Professor of Electrical Engineering and director of the Center for Language and Speech Processing.

Jelinek, a native of the Czech Republic, earned his bachelor’s, master’s and doctoral degrees, all in electrical engineering, from MIT.

After obtaining his doctorate in 1962, Jelinek joined the faculty of Cornell University, where he continued to study information theory. A decade later, he applied for a summer position at the IBM T.J. Watson Research Center. Soon he was appointed to head the center’s large Continuous Speech Recognition group on a full-time basis.

“Our group completely revolutionized the standard approach to speech recognition” Jelinek said in a 2001 speech. “This was made easy by the fact that practically none of us was educated in any subject related to speech. Those who contributed most had their doctorates in information theory or physics. Their creativity, determination, originality, and courage to risk was what accounted for our success.”

Jelinek’s son, William Jelinek, said his father “was extremely proud of the speech recognition group he led at IBM for 21 years. When he retired from IBM in 1993, he wanted to continue to do research, and Johns Hopkins gave him the opportunity when the school asked him to direct the Center for Language and Speech Processing there.”

At Johns Hopkins, his colleagues said Jelinek was able to build on his ground-breaking work at IBM and share it with students and other engineering faculty members. He led highly regarded summer workshops that brought together speech and language processing researchers from industry, government, and academia, along with undergraduates and graduate students.

SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues