Translation as Cross-Domain Knowledge: Attention Augmentation for Unsupervised Cross-Domain Segmenting and Labeling Tasks ( Findings of EMNLP 2021)
We propose a novel paradigm based on attention augmentation to introduce crucial cross-domain knowledge via a translation system. The proposed paradigm enables the model attention to draw cross-domain knowledge indicated by the implicit word-level cross-lingual alignment between the input and its corresponding translation. Aside from the model requiring crosslingual input, we also establish an off-the-shelf model which eludes the dependency on crosslingual translations. Experiments demonstrate that our proposal significantly advances the state-of-the-art results of cross-domain Chinese segmenting and labeling tasks.
We propose a simple yet effective algorithm named FSSD based on a novel observation: in a trained neural network, OoD samples with bounded norms well concentrate in the feature space. It achieves state-of-the-art performance on various OoD detection benchmark and enjoys robustness to slight corruption in test data.