Papers for

legal and compliance teams

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

New system evaluates explainability methods for document AI models

DocXEval: A Systematic Evaluation of Explainability in Multimodal Document Understanding

Abstract: Multimodal transformer models have driven unprecedented performance breakthroughs in modern document understanding. However, the inherent black-box nature of these models remains a major obstacle to their safe deployment in high-stakes sectors. Despite significant progress in explainability for vision and language tasks, explainable AI (XAI) research tailored to multimodal document understanding has received little attention. To address this gap, we present \ourframework{}, the first comprehensive explainability framework that adapts existing XAI methods and evaluation metrics to the multimodal (text, layout, and image) document understanding context, while introducing efficient multi-target adaptations of both methods and metrics for domain-specific tasks such as semantic entity recognition. Using \ourframework{}, we systematically evaluate 13 XAI methods across 3 document understanding tasks, 8 datasets, 4 models, and 17 evaluation metrics, and find that only a few methods perform consistently across all evaluation criteria, with Occlusion emerging as the most faithful and interpretable method overall. Through a dedicated meta-evaluation, we observe moderate to strong agreement among several faithfulness and complexity metrics while identifying certain ones that produce unreliable results. Finally, through multiple dataset-wide diagnostic analyses, we uncover key insights into model behavior, including the contribution of individual modalities across tasks and labels, spatial attribution patterns around target entities, and potential dataset biases. Overall, this work serves as both a benchmark and a practical guide for future XAI research in multimodal document understanding. To support reproducibility, we publicly release our code at https://github.com/saifullah3396/docxeval/.

Tue 27 OctComputation and LanguageComputer Vision and Pattern Recognition
The gist
Modern AI models that understand documents mix images, text, and layout information but are hard to explain, especially for important uses. The authors created a framework called DocXEval that tests many explainability methods specifically for these multimodal document models. They found that only a few methods work well consistently, with one called Occlusion standing out. Their work also shows how much each part of the document helps the model and points out dataset quirks, providing a useful guide for improving explainability in these systems.
Open → 2610.36929v1