Lawrence R. Rabiner (born 28 September 1943 in Brooklyn, New York) is an electrical engineer working in the fields of digital signal processing and speech processing; in particular in digital signal processing for automatic speech recognition. He has worked on systems for AT&T Corporation for speech recognition.
He holds a joint academic appointment between Rutgers University and the University of California, Santa Barbara.
Lawrence Rabiner was born in Brooklyn, NY in 1943. During his studies at MIT, he participated in the cooperative program at AT&T Bell Laboratories, during which he worked on digital circuit design and binaural hearing. After obtaining his PhD in 1967, he joined AT&T Bell Laboratories' research division in Murray Hill, NJ as a Member of Technical Staff. He was promoted to Supervisor in 1972, Department Head in 1985, Director in 1990, and Functional Vice-President in 1995. He joined the newly created AT&T Labs - Research in 1996 as Director of the Speech and Image Processing Services Research Laboratory. He was promoted Vice-President of Research in 1998, succeeding Sandy Fraser, where he managed broad programs in communication, computing, and information sciences. He retired from AT&T in 2002 and joined the department of Electrical Engineering at Rutgers University, with a joint appointment at the University of California in Santa Barbara.
Dr. Rabiner has pioneered a range of novel algorithms for digital filtering and digital spectrum analysis. The most well known of these algorithms are the Chirp z-Transform method (CZT) of spectral analysis, a range of optimal FIR (finite impulse response) digital filter design methods based on linear programming and Chebyshev approximation methods, and a class of decimation/interpolation methods for digital sampling rate conversion. In the area of speech processing, Dr. Rabiner has made contributions to the fields of pitch detection, speech synthesis and speech recognition. Dr. Rabiner built one of the first digital speech synthesizers that was able to convert arbitrary text to intelligible speech. In the area of speech recognition, Dr. Rabiner was a major contributor to the creation of the statistical method of representing speech that is known as hidden Markov modeling (HMM). Dr. Rabiner was the first to publish the scaling algorithm for the Forward-Backward method of training of HMM recognizers. His research showed how to successfully implement an HMM system based on either discrete or continuous density parameter distributions. His tutorial paper on HMM is highly cited. Dr. Rabiner’s research resulted in a series of speech recognition systems that went into deployment by AT&T to enable automation of a range of ‘operator services’ that previously had been carried out using live operators. One such system, called the Voice Recognition Call Processing (VRCP) system, automated a small vocabulary recognition system (5 active words) with word spotting and barge-in capability. It resulted in savings of several hundred millions of dollars annually for AT&T.