Data science and signal processing

Data science resides at the nexus of the physical world and computation—where fundamental sciences of systems, networks, and communications enable the design of technologies and algorithms that extract information from large amounts of unstructured data produced by physical sensors and systems, discover patterns, and make predictions and critical decisions using a variety of tools including machine learning and neural networks, while guaranteeing reliability and robustness, and preserving security and privacy.

The interaction of data science and technology with the world is via signal processing: detecting, transcoding, understanding and generating time-dependent and space-dependent signals in the broadest sense. This includes signals in optical, electrical, acoustic, chemical, biological, textual, and social media. Signal processing is a discipline of applied mathematics, using the tools of information theory, probability and statistics, vector spaces, harmonic analysis, optimization, and machine learning. To quote from the IEEE Signal Processing Society website, “Signal processing is at the heart of our modern world, powering today's entertainment and tomorrow's technology. It's at the intersection of biotechnology and social interactions. It enhances our ability to communicate and share information. Signal processing is the science behind our digital lives.”

Current research projects include processing of speech, audio, image, video, genomic, and social network signals, computational imaging, human-computer intelligent interaction, and visual analytics in domains including geospatial, social networking, free-field audio, bioelectric, and biomedical imaging.

Faculty with primary interest in this area

Affiliate faculty with primary interest in this area

Faculty with secondary interest in this area