Moqian Tian
Graduate Student

Spatiotemporal Learning of Novel Objects

How do we recognize objects despite the vast changes in their appearance due to changes in viewpoint, size, position and lighting
conditions? Previous research has demonstrated that experience may play an important role in generating invariant object recognition. I use psychophysical methods to measure the effects of spatiotemporal unsupervised learning in generating viewpoint invariant representations of novel objects. Specifically, I ask whether there is enhanced learning when objects are learned in a spatiotemporally correlated fashion compared to random learning sequences.