Semi-supervised domain adaptation (SSDA) is a novel branch of machine learning that scarce labeled target examples are available, compared with unsupervised domain adaptation. To make effective use of these additional data so as to bridge the domain gap, one possible way is to generate adversarial examples, which are images with additional perturbations, between the two domains and fill the domain gap. Adversarial training has been proven to be a powerful method for this purpose. However, the traditional adversarial training adds noises in arbitrary directions, which is inefficient to migrate between domains, or generate directional noises from the source to target domain and reverse. In this work, we devise a general bidirectional adversarial training method and employ gradient to guide adversarial examples across the domain gap, i.e., the Adaptive Adversarial Training (AAT) for source to target domain and Entropy-penalized Virtual Adversarial Training (E-VAT) for target to source domain. Particularly, we devise a Bidirectional Adversarial Training (BiAT) network to perform diverse adversarial trainings jointly. We evaluate the effectiveness of BiAT on three benchmark datasets and experimental results demonstrate the proposed method achieves the state-of-the-art.

The Xining Basin lies in the transitional zone between the arid Asian interior and the East Asian monsoon region. The continuous Cenozoic sediments in the basin provide a unique archive recording the uplift of the Tibetan Plateau and its environmental effects on central Asian aridification and Asian monsoon evolution. However, sediments deposited since the middle Middle Miocene have not been precisely dated, hindering our ability to address these issues. Here we dated a 336 m‐thick section containing many Late Miocene fossil mammals from the eastern basin. High‐resolution paleomagnetism revealed 16 normal and 16 reversed zones that correlate well with chrons 3n to 5Ar.1r of the Geomagnetic Polarity Time Scale, constraining the section to ∼12.7–4.8 Ma. The changes in lithofacies from floodplain to braided river at ∼8.6 Ma and to thick alluvial fan at ∼6.3 Ma with predominantly southerly paleocurrent directions occur simultaneously with an increase in the sedimentation rates, representing two periods of rapid uplift in the eastern Qilian Shan to the north. Our results provide a robust oldest age constraint (<4.8 Ma) on the ancient Huang Shui and Yellow River terraces on the basin sediments and eolian deposits on the terraces. The presence of well‐developed floodplain paleosols, shallow lake marls, and many large mammals in the lower section suggests that the monsoon climate in the late Middle Miocene to Late Miocene was more humid and stronger than that at present.

No abstract