Frederick Jelinek facts for kids

Frederick Jelinek (born November 18, 1932 – died September 14, 2010) was a Czech-American scientist. He was a leader in fields like information theory, automatic speech recognition, and natural language processing. He is famous for saying, "Every time I fire a linguist, the performance of the speech recognizer goes up." This quote highlights his belief in using math and statistics for language technology.

Jelinek was born in Czechoslovakia before World War II. His family moved to the United States when the communist government took over. He studied engineering at the Massachusetts Institute of Technology (MIT). After teaching at Cornell University for ten years, he joined IBM Research. In 1961, he married Czech screenwriter Milena Jelinek. At IBM, his team made big steps in computer speech recognition and machine translation. Later, he led the Center for Language and Speech Processing at Johns Hopkins University for 17 years. He was still working there when he passed away.

Early Life and Education

Frederick Jelinek was born Bedřich Jelínek on November 18, 1932, in Kladno, which is now in the Czech Republic. His parents were Vilém and Trude Jelínek. His father was Jewish. His mother was born in Switzerland and had converted to Judaism. His father, a dentist, had planned for the family to escape to England before Nazi Germany took over. However, he decided to stay and later died in a concentration camp in 1945.

In 1941, Frederick, his sister, and his mother had to move to Prague. Luckily, his mother's background helped them avoid being sent to concentration camps. After the war, Frederick went to a gymnasium (a type of high school). He had missed several years of school because Jewish children were not allowed to be educated after 1942. His mother worked hard to help them move to the U.S. so he could get a good education.

Jelinek studied engineering in evening classes at the City College of New York. He also received scholarships that allowed him to study at the Massachusetts Institute of Technology (MIT). He earned his Ph.D. in 1962. He chose to study electrical engineering because it included "the theory of information," which he found fascinating.

Family and Career Journey

In 1957, Jelinek visited Prague and met Milena Tobolová, a film student. Her screenplay was the basis for a movie called Easy Life. He proposed to her over the phone during a stopover on his flight back to the U.S. Milena faced difficulties from the authorities in Czechoslovakia. With help from important people like Jerome Wiesner and Cyrus S. Eaton, she was allowed to move to the U.S. in January 1961.

After finishing his studies, Jelinek became interested in linguistics, the study of language. He planned to work at Cornell University, but it didn't work out. For the next ten years, he focused on information theory. He had worked at IBM during a break from Cornell. In 1972, he started working full-time at IBM. He stayed there for over twenty years. He became the head of the Continuous Speech Recognition group at the Thomas J. Watson Research Center. Even though his team had great success, Jelinek's work was not well-known in his home country. This was because Czech scientists were not allowed to attend important conferences.

After communism ended in 1989, Jelinek helped build scientific connections between the U.S. and the Czech Republic. He visited often to give lectures. He also helped convince IBM to set up a computing center at Charles University. In 1993, he left IBM and joined Johns Hopkins University. He became the director of the Center for Language and Speech Processing. He was still working there when he died from a heart attack in September 2010. He was survived by his wife, children, sister, and grandchildren.

Amazing Work in Language Technology

In the 1950s, Information theory was a very popular scientific field. However, some experts worried that people were using it too broadly. For example, applying it to natural language processing (NLP) problems, like machine translation, faced challenges. One big reason was a book by Noam Chomsky in 1957. He argued that math-based models didn't help understand how language structure works. This idea fit well with the thinking in artificial intelligence at the time, which preferred rule-based systems.

Another challenge was a report in 1966 that suggested the government stop funding machine translation research. This report claimed the field was full of "mad inventors." These factors almost stopped research in this area for a while.

Jelinek became interested in linguistics after his wife started studying it at MIT. He often went to lectures and even thought about changing his own field of study. When he started at IBM, his team took a very different approach. While most other researchers worked on small, specific problems, Jelinek led IBM's effort to solve the bigger problem of general speech recognition.

Jelinek saw speech recognition as a problem of information theory. He thought of it like a "noisy channel," where the sound signal was unclear. This was a bold idea at the time. His team introduced the idea of "perplexity" in their first model, published in 1976. This model helped turn the speech recognition problem into one of creating two statistical models. Their next model, called Tangora, used n-grams (like trigrams), which look at groups of words. Even though this model seemed simple, it was very effective. The same trigram approach was used for individual sounds in words.

The step-by-step research methods developed at IBM became very important in the field. In the mid-1980s, a government agency called DARPA started funding NLP research again. They required teams to use similar methods, share data, and have clear ways to measure success. Jelinek's team's research needed huge amounts of data to train their computer programs. This led to the creation of the Linguistic Data Consortium, which provides language data for research.

In the 1980s, Jelinek's team also tried to use their methods for other problems, like machine translation and predicting stock prices. Many of the methods developed for speech recognition are now used in most machine translation systems today. Some experts say that Jelinek's work showed that engineering success can lead to new scientific discoveries, rather than the other way around.

Awards and Recognition

Frederick Jelinek received many awards for his important work. His papers won "best paper" awards several times. He also received company awards while at IBM.

Some of his major awards include:

• The Society Award from the IEEE Signal Processing Society in 1997.
• The ESCA Medal for Scientific Achievement in 1999.
• The James L. Flanagan Speech and Audio Processing Award in 2005.
• The Lifetime Achievement Award from the Association for Computational Linguistics in 2009.

He also received an honorary Ph.D. from Charles University in 2001. In 2006, he was elected to the National Academy of Engineering, which is a very high honor. In 2008, he became one of the first twelve fellows of the International Speech Communication Association.