Description
In addition to communication, hearing is useful for informing us about general events in the world, since any kind of physical contact or abrasion will radiate sound. For computers, however, the automatic detection and recognition of sounds other than speech and music has received relatively little attention. I will briefly review past work in automatic recognition of these so-called environmental sounds, including several recent evaluations.
The current generation of big-data-fueled machine learning classifiers offers the possibility of huge advances over the current state of the art, but demands large numbers of training examples. I'll discuss various responses to this need, look at some results, and consider likely future outcomes.
Speaker Bio
Daniel P. W. Ellis received the Ph.D. degree in electrical engineering from MIT, where he was a Research Assistant in the Machine Listening Group of the Media Lab. He spent several years as a Research Scientist at the International Computer Science Institute, Berkeley, CA. He joined Columbia University in 2000 and set up the Laboratory for Recognition and Organization of Speech and Audio (LabROSA). Currently, he is on leave from Columbia while working with the Sound Understanding group at Google in New York.