What is Flederick Jarinik's expert?

Flederick Jarinik is an expert in speech recognition and natural language processing.

Jerry Nick was born in 1932. His family is a wealthy Jewish family in the Czech Republic and could have received a good education. However, these were all broken by World War II. My father died in a concentration camp, but Jarinik wandered the streets all day, returned to school after the war, and soon caught up with his classmates, but he didn't get an A either.

1949, her mother immigrated to America, and she was very poor. She had to work to support her family, so her studies were naturally delayed. However, his idea of working hard for his ideal has always been firm. At first, he wanted to be a lawyer, defending lawbreakers like his father. However, it was soon discovered that fluent English would make him suffer in court.

He was determined to become a doctor again, but was turned away by the high tuition of Harvard Medical School. Just when MIT offered him a full scholarship, he began to study electrical engineering. At MIT, I met Dr. Shannon, the originator of information theory, and Jagerson, the master of linguistics. These masters have a great influence on him.

Jarinik received his Ph.D. from MIT, taught at Cornell University, and then went to IBM, where he established an unprecedented powerful research team and proposed a framework for statistical language recognition. Their BCJR algorithm is listed by IBM as one of IBM's greatest contributions to mankind.

Jarinik once studied the problem of data communication with Professor Lloyd elwood Shannon, the founder of information theory. He was the best communication expert of his time. So when he faced the problem of speech recognition, he naturally thought that speech recognition can actually be compared to a communication problem.

The essence of speech recognition is that the "speaker" wants the "listener" to understand his own ideas. This process can be divided into three steps:

The speaker organizes the ideas in his mind into language and expresses them through words; The sound is transmitted to the listener, and changes from a sound signal to a nerve signal in the listener's brain; The listener converts neural signals into the speaker's thoughts.

This process is actually equivalent to "coding-transmission-decoding" in the communication process. Therefore, Jarinik thought of borrowing the idea of solving communication problems, regarding the inaccurate speech recognition as noise interference in the communication process, and using big data to solve the interference problem to solve the speech recognition problem.