Joint Hierarchical Domain Adaptation and Feature Learning

Complex visual data contain discriminative structures that are difficult to be fully captured by any single feature descriptor. While recent work in domain adaptation focuses on adapting a single hand-crafted feature, it is important to perform adaptation on a hierarchy of features to exploit the richness of visual data. We propose a novel framework for domain adaptation using a sparse and hierarchical network (DASH-N). Our method jointly learns a hierarchy of features together with transformations that address the mismatch between different domains. The building block of DASH-N is designed using the theory of latent sparse representation. It employs a dimensionality reduction step that can prevent the data dimension from increasing too fast as we increase the depth of the hierarchy. Experimental results show that our method compares favorably with competing state-of-the-art methods. Moreover, it is shown that a multi-layer DASH-N performs better than a single-layer DASH-N.