Test-Time Adaptation for Visual Document Understanding

Published: 10 Jun 2023, Last Modified: 10 Jun 2023Accepted by TMLREveryoneRevisionsBibTeX
Abstract: For visual document understanding (VDU), self-supervised pretraining has been shown to successfully generate transferable representations, yet, effective adaptation of such representations to distribution shifts at test-time remains to be an unexplored area. We propose DocTTA, a novel test-time adaptation method for documents, that does source-free domain adaptation using unlabeled target document data. DocTTA leverages cross-modality self-supervised learning via masked visual language modeling, as well as pseudo labeling to adapt models learned on a \textit{source} domain to an unlabeled \textit{target} domain at test time. We introduce new benchmarks using existing public datasets for various VDU tasks, including entity recognition, key-value extraction, and document visual question answering. DocTTA shows significant improvements on these compared to the source model performance, up to 1.89\% in (F1 score), 3.43\% (F1 score), and 17.68\% (ANLS score), respectively.
Submission Length: Regular submission (no more than 12 pages of main content)
Previous TMLR Submission Url: https://openreview.net/forum?id=w6eERU7wql&referrer=%5BAuthor%20Console%5D(%2Fgroup%3Fid%3DTMLR%2FAuthors%23your-submissions)
Changes Since Last Submission: It was desk-rejected because of a wrong font type (times). We fixed it to comply with TMLR template.
Video: https://saynaebrahimi.github.io/index_files/doctta/DocTTA_TMLR.mp4
Supplementary Material: zip
Assigned Action Editor: ~Yale_Song1
License: Creative Commons Attribution 4.0 International (CC BY 4.0)
Submission Number: 988