Bridging the Theoretical Bound and Deep Algorithms for Open Set Domain Adaptation
Li Zhong,Zhen Fang,Feng Liu,Bo Yuan,Guangquan Zhang,Jie Lu
DOI: https://doi.org/10.1109/tnnls.2021.3119965
IF: 14.255
2021-01-01
IEEE Transactions on Neural Networks and Learning Systems
Abstract:In the unsupervised open set domain adaptation (UOSDA), the target domain contains unknown classes that are not observed in the source domain. Researchers in this area aim to train a classifier to accurately: 1) recognize unknown target data (data with unknown classes) and 2) classify other target data. To achieve this aim, a previous study has proven an upper bound of the target-domain risk, and the open set difference, as an important term in the upper bound, is used to measure the risk on unknown target data. By minimizing the upper bound, a shallow classifier can be trained to achieve the aim. However, if the classifier is very flexible [e.g., deep neural networks (DNNs)], the open set difference will converge to a negative value when minimizing the upper bound, which causes an issue where most target data are recognized as unknown data. To address this issue, we propose a new upper bound of target-domain risk for UOSDA, which includes four terms: source-domain risk, ϵ -open set difference ( ∆<sub>ϵ</sub> ), distributional discrepancy between domains, and a constant. Compared with the open set difference, ∆<sub>ϵ</sub> is more robust against the issue when it is being minimized, and thus we are able to use very flexible classifiers (i.e., DNNs). Then, we propose a new principle-guided deep UOSDA method that trains DNNs via minimizing the new upper bound. Specifically, source-domain risk and ∆<sub>ϵ</sub> are minimized by gradient descent, and the distributional discrepancy is minimized via a novel open set conditional adversarial training strategy. Finally, compared with the existing shallow and deep UOSDA methods, our method shows the state-of-the-art performance on several benchmark datasets, including digit recognition [modified National Institute of Standards and Technology database (MNIST), the Street View House Number (SVHN), U.S. Postal Service (USPS)], object recognition (Office-31, Office-Home), and face recognition [pose, illumination, and expression (PIE)].
computer science, artificial intelligence, theory & methods,engineering, electrical & electronic, hardware & architecture