Lawrence K. Saul
Professor, Computer Science and Engineering
Machine learning, pattern recognition, voice processing, auditory computation and methods for high dimensional data analysis. Lawrence Saul is internationally known for his work on high dimensional data analysis and visualization. Combining ideas from the fields of computer science and statistics, he has developed powerful, new algorithms for revealing low dimensional structure in high dimensional data. Though capable of identifying complex nonlinear relationships, his approaches retain the tractability of traditional linear methods. His work on nonlinear dimensionality reduction has applications in many areas of science and engineering, including computational neuroscience, pattern recognition, and information processing in sensor networks.Saul is also an expert in the application of ideas from machine learning to problems in audio processing. With his students, he is focused on improved acoustic models for automatic speech recognition, efficient (real-time) algorithms for auditory scene analysis, and robust integration of cues in different frequency bands. Much of his work in speech and audio processing is inspired by psychoacoustic models of human listeners. He is especially interested in "the cocktail party" problem -- how to follow a single voice in a room with multiple overlapping speakers. Algorithms for solving this problem have widespread applications in audio surveillance and information retrieval.
Jacobs School Faculty Update Your Profile