Making Document-Level Information Extraction Right for the Right Reasons
This addresses the need for more interpretable and auditable models in domains like healthcare, though it is incremental in improving existing methods.
The paper tackles the problem of document-level information extraction models learning spurious correlations, proposing a predict-select-verify framework with evidence supervision to improve plausibility without sacrificing accuracy, as demonstrated on brain MRI reports and a modified DocRED dataset.
Document-level models for information extraction tasks like slot-filling are flexible: they can be applied to settings where information is not necessarily localized in a single sentence. For example, key features of a diagnosis in a radiology report may not be explicitly stated in one place, but nevertheless can be inferred from parts of the report's text. However, these models can easily learn spurious correlations between labels and irrelevant information. This work studies how to ensure that these models make correct inferences from complex text and make those inferences in an auditable way: beyond just being right, are these models "right for the right reasons?" We experiment with post-hoc evidence extraction in a predict-select-verify framework using feature attribution techniques. We show that regularization with small amounts of evidence supervision during training can substantially improve the quality of extracted evidence. We evaluate on two domains: a small-scale labeled dataset of brain MRI reports and a large-scale modified version of DocRED (Yao et al., 2019) and show that models' plausibility can be improved with no loss in accuracy.