Batch normalization (BN) is widely used in modern deep neural networks, which has been shown to represent the domain-related knowledge, and thus ineffective for cross-domain tasks like unsupervised domain adaptation (UDA). Existing BN variant methods aggregate source target knowledge same channel module. However, misalignment between features of corresponding channels across domains often leads...

D-instanton amplitudes suffer from various infrared divergences associated with tachyonic or massless open string modes, leading to ambiguous contribution amplitudes. It has been shown previously that field theory can resolve these ambiguities and lead unambiguous expressions for contributions amplitudes, except an overall normalization constant remains undetermined. In this paper we show theor...