Results 11 to 20 of about 9,581,242 (372)
A retrieval and ranking method of mathematical documents based on CA-YOLOv5 and HFS
Mathematical Biosciences and Engineering, 2022 In a retrieval system for mathematical documents based on mathematical expressions, the input and matching of mathematical expressions are key steps that affect the system's usability, accessibility and efficiency because of their special attributes ...
Xinpeng Xu, Xuedong Tian, Fang Yang
doaj +1 more source
DocVQA: A Dataset for VQA on Document Images [PDF]
IEEE Workshop/Winter Conference on Applications of Computer Vision, 2020 We present a new dataset for Visual Question Answering (VQA) on document images called DocVQA. The dataset consists of 50,000 questions defined on 12,000+ document images.
Minesh Mathew +3 more
semanticscholar +1 more source
New Trends in Improving Public Service Delivery in Ukraine
Науковий вісник Мукачівського державного університету. Серія Економіка, 2021 In the aspect of the European integration aspirations of Ukrainian society and the social development of the nation state, the issues of its service function formation in the form of public services are becoming more and more relevant in Ukraine.
Tymur Slobodeniuk
doaj +1 more source
Cross-Depicted Historical Motif Categorization and Retrieval with Deep Learning
Journal of Imaging, 2020 In this paper, we tackle the problem of categorizing and identifying cross-depicted historical motifs using recent deep learning techniques, with aim of developing a content-based image retrieval system. As cross-depiction, we understand the problem that
Vinaychandran Pondenkandath +4 more
doaj +1 more source
SPECTER: Document-level Representation Learning using Citation-informed Transformers [PDF]
Annual Meeting of the Association for Computational Linguistics, 2020 Representation learning is a critical ingredient for natural language processing systems. Recent Transformer language models like BERT learn powerful textual representations, but these models are targeted towards token- and sentence-level training ...
Arman Cohan +4 more
semanticscholar +1 more source
Syntax of Native Advertising Publications in Glossy Periodicals
Научный диалог, 2020 The question of the originality of the syntax of native advertising publications in Russian editions of international glossy magazines in 2018-2020 is considered. The relevance of the chosen topic is due to the interest of domestic and foreign experts in
O. A. Selemeneva
doaj +1 more source
LayoutLMv2: Multi-modal Pre-training for Visually-rich Document Understanding [PDF]
Annual Meeting of the Association for Computational Linguistics, 2020 Pre-training of text and layout has proved effective in a variety of visually-rich document understanding tasks due to its effective model architecture and the advantage of large-scale unlabeled scanned/digital-born documents.
Yang Xu +11 more
semanticscholar +1 more source
Document Ranking with a Pretrained Sequence-to-Sequence Model [PDF]
Findings, 2020 This work proposes the use of a pretrained sequence-to-sequence model for document ranking. Our approach is fundamentally different from a commonly adopted classification-based formulation based on encoder-only pretrained transformer architectures such ...
Rodrigo Nogueira +3 more
semanticscholar +1 more source
Conducting a Qualitative Document Analysis
The Qualitative Report, 2022 Document analysis has been an underused approach to qualitative research. This approach can be valuable for various reasons. When used to analyze pre-existing texts, this method allows researchers to conduct studies they might otherwise not be able to ...
H. Morgan
semanticscholar +1 more source
Unifying Vision, Text, and Layout for Universal Document Processing [PDF]
Computer Vision and Pattern Recognition, 2022 We propose Universal Document Processing (UDOP), a foundation Document AI model which unifies text, image, and layout modalities together with varied task formats, including document understanding and generation.
Zineng Tang +8 more
semanticscholar +1 more source