I am currently a postdoctoral fellow at the University of Southern California in the Laboratory for Neural Computation working with Prof. Bartlett Mel on visual object recognition. Our goal is to devise learning rules that can develop a feature-detector hierarchy similar to that proposed by Fukushima et al. (1983) in order to recognize objects independent of location, scale, or orientation. We have analyzed the tradeoff of feature complexity and feature detector invariance to understand the kind of features we want to learn, and we are currently trying to develop learning principles that will produce such features.