This information is updated when plans for the monthly meeting are finalized.

Automatic Speech Recognition

Edward C. Bronson

AT&T Bell Laboratories

Tuesday March 28th at 7:00pm College of Dupage Room K-157


Conversing with a computer using natural language has been the dream of researchers and science fiction fen for decades. Over the past few years, the maturity of speech recognition algorithms and the increased processing power available from lower-cost digital signal processing ICs has made Automatic Speech Recognition (ASR) a viable reality for many commercial applications.

An overview of the state of the art in computer speech processing will be presented. Computer algorithms commonly used in speech recognition will be discussed including such topics as filtering, cepstral feature extraction, dynamic time warping, and hidden Markov models. Details of speech model building necessary for successful ASR along with the processing techniques required to make speech recognition operate in a real world user environment will also be presented. The talk will include a discussion of the issues relating to the hardware platforms required to perform recognition in real-time with a special emphasis on current and future ASR applications being employed in the telephone network.

Speaker Bio

Edward Bronson is a Member of Technical Staff in the Speech Processing Departmemt of AT&T Bell Laboratories in Naperville, IL. He received his Ph.D. in Electrical Engineering at Purdue University concentrating on parallel computer architectures for real-time speech understanding. He has authored dozens of research papers on speech algorithms, computer architecture design, simulation, and, more recently, on pattern recognition in molecular biology.

Getting to the meeting

Directions on how to get to the meeting include directions on getting to the College of DuPage (COD) from Interstate-88, locating the correct building on campus, and the location within the building.

For more information about UniForum Chicago contact

UniForum is a registered trademark of UniForum Association