Instance-Based Neural Dependency Parsing

Abstract

Interpretable rationales for model predictions are crucial in practical applications. We develop neural models that possess an interpretable inference process for dependency parsing. Our models adopt instance-based inference, where dependency edges are extracted and labeled by comparing them to edges in a training set. The training edges are explicitly used for the predictions; thus, it is easy to grasp the contribution of each edge to the predictions. Our experiments show that our instance-based models achieve competitive accuracy with standard neural models and have the reasonable plausibility of instance-based explanations.

Span-based Named Entity Recognition by Generating and Compressing Information

“…Results in Table 3 show that on the GENIA corpus, the joint model could perform better than (1) the baseline and (2) the two span-based models: the instanced-based NER model (Ouchi et al, 2021) and the boundary-aware one (Zheng et al, 2019). However, our model produced lower F1 scores than BENSC (Tan et al, 2020) and MHSA (Xu et al, 2021).…”