Discrepancy Directed Model Acquisition for Adaptive Perceptual Systems

For complex perceptual tasks that are char acterized by object occlusion and non stationarity recognition systems with adaptive signal processing front ends have been devel oped These systems rely on hand crafted sym bolic object models which constitutes a knowl edge acquisition bottleneck We propose an ap proach to automate object model acquisition that relies on the detection and resolution of signal processing and interpretation discrepan cies The approach is applied to the task of acquiring acoustic event models for the Sound Understanding Testbed SUT