Xu Zhong
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Xu Zhong has a total of 47 co-authors. Here are the top co-authors by number of shared publications:
- Matt Duckham
RMIT University
5 shared publications - Antonio Jimeno Yepes
IBM Research Australia
4 shared publications Bill Moran
4 shared publicationsNicola Cooley
4 shared publications- Peter M. Farrell
University of Melbourne
4 shared publications Xuezhi Wang
3 shared publications
Latest Publications
- 15 July 2022
- Published by SPIE-Intl Soc Optical Eng
- Vol. 12258, 92-569
Research on 3D visualization modeling and lightweight display technology of transmission channel
Distribution network communication operation and maintenance platform rely on communication network. The distribution network substation needs to obtain distribution network operation data configuration through communication network. The reliability of communication network directly affects the protection performance. Therefore, there is a need to establish a communication network that is reliable and meet protection performance requirements. In order to improve the engineering application of distribution network communication operation and maintenance platform, it is necessary to quantitatively evaluate the data quality of the distribution network communication operation and maintenance platform. Based on the construction background of smart distribution network, this paper studies the architecture of distribution network communication operation and maintenance platform. Starting from the requirements of information interconnection, intercommunication, interoperability and plug-and-play, the description specification of distribution network communication operation is established according to the principle of artificial intelligence data analysis. The unified modeling of protection and control system is realized. By analyzing the business composition of regional protection and self-healing control system of distribution network, the size of network information flow is quantitatively evaluated, which provides a basis for the engineering application of regional protection and self-healing control system of the distribution network. Experimental analysis shows that compared with the traditional method, the comprehensive performance of this method is improved by 10%. The partial performance is improved by 40%, which can support the master-slave regional protection and control business system.
- 15 July 2022
- Published by SPIE-Intl Soc Optical Eng
- Vol. 12258, 92-569
Research on 3D visualization modeling and lightweight display technology of transmission channel
Distribution network communication operation and maintenance platform rely on communication network. The distribution network substation needs to obtain distribution network operation data configuration through communication network. The reliability of communication network directly affects the protection performance. Therefore, there is a need to establish a communication network that is reliable and meet protection performance requirements. In order to improve the engineering application of distribution network communication operation and maintenance platform, it is necessary to quantitatively evaluate the data quality of the distribution network communication operation and maintenance platform. Based on the construction background of smart distribution network, this paper studies the architecture of distribution network communication operation and maintenance platform. Starting from the requirements of information interconnection, intercommunication, interoperability and plug-and-play, the description specification of distribution network communication operation is established according to the principle of artificial intelligence data analysis. The unified modeling of protection and control system is realized. By analyzing the business composition of regional protection and self-healing control system of distribution network, the size of network information flow is quantitatively evaluated, which provides a basis for the engineering application of regional protection and self-healing control system of the distribution network. Experimental analysis shows that compared with the traditional method, the comprehensive performance of this method is improved by 10%. The partial performance is improved by 40%, which can support the master-slave regional protection and control business system.
- Open Access
Global Table Extractor (GTE): A Framework for Joint Table Identification and Cell Structure Recognition Using Visual Context
Documents are often used for knowledge sharing and preservation in business and science, within which are tables that capture most of the critical data. Unfortunately, most documents are stored and distributed as PDF or scanned images, which fail to preserve logical table structure. Recent vision-based deep learning approaches have been proposed to address this gap, but most still cannot achieve state-of-the-art results. We present Global Table Extractor (GTE), a vision-guided systematic framework for joint table detection and cell structured recognition, which could be built on top of any object detection model. With GTE-Table, we invent a new penalty based on the natural cell containment constraint of tables to train our table network aided by cell location predictions. GTE-Cell is a new hierarchical cell detection network that leverages table styles. Further, we design a method to automatically label table and cell structure in existing documents to cheaply create a large corpus of training and test data. We use this to enhance PubTabNet with cell labels and create FinTabNet, real-world and complex scientific and financial datasets with detailed table structure annotations to help train and test structure recognition. Our framework surpasses previous state-of-the-art results on the ICDAR 2013 and ICDAR 2019 table competition in both table detection and cell structure recognition with a significant 5.8% improvement in the full table extraction system. Further experiments demonstrate a greater than 45% improvement in cell structure recognition when compared to a vanilla RetinaNet object detection model in our new out-of-domain FinTabNet.
Open AccessGlobal Table Extractor (GTE): A Framework for Joint Table Identification and Cell Structure Recognition Using Visual Context
Documents are often used for knowledge sharing and preservation in business and science, within which are tables that capture most of the critical data. Unfortunately, most documents are stored and distributed as PDF or scanned images, which fail to preserve logical table structure. Recent vision-based deep learning approaches have been proposed to address this gap, but most still cannot achieve state-of-the-art results. We present Global Table Extractor (GTE), a vision-guided systematic framework for joint table detection and cell structured recognition, which could be built on top of any object detection model. With GTE-Table, we invent a new penalty based on the natural cell containment constraint of tables to train our table network aided by cell location predictions. GTE-Cell is a new hierarchical cell detection network that leverages table styles. Further, we design a method to automatically label table and cell structure in existing documents to cheaply create a large corpus of training and test data. We use this to enhance PubTabNet with cell labels and create FinTabNet, real-world and complex scientific and financial datasets with detailed table structure annotations to help train and test structure recognition. Our framework surpasses previous state-of-the-art results on the ICDAR 2013 and ICDAR 2019 table competition in both table detection and cell structure recognition with a significant 5.8% improvement in the full table extraction system. Further experiments demonstrate a greater than 45% improvement in cell structure recognition when compared to a vanilla RetinaNet object detection model in our new out-of-domain FinTabNet.
- Open Access
Forget Me Not: Reducing Catastrophic Forgetting for Domain Adaptation in Reading Comprehension
The creation of large-scale open domain reading comprehension data sets in recent years has enabled the development of end-to-end neural comprehension models with promising results. To use these models for domains with limited training data, one of the most effective approach is to first pretrain them on large out-of-domain source data and then fine-tune them with the limited target data. The caveat of this is that after fine-tuning the comprehension models tend to perform poorly in the source domain, a phenomenon known as catastrophic forgetting. In this paper, we explore methods that overcome catastrophic forgetting during fine-tuning without assuming access to data from the source domain. We introduce new auxiliary penalty terms and observe the best performance when a combination of auxiliary penalty terms is used to regularise the fine-tuning process for adapting comprehension models. To test our methods, we develop and release 6 narrow domain data sets that could potentially be used as reading comprehension benchmarks.
Open AccessForget Me Not: Reducing Catastrophic Forgetting for Domain Adaptation in Reading Comprehension
The creation of large-scale open domain reading comprehension data sets in recent years has enabled the development of end-to-end neural comprehension models with promising results. To use these models for domains with limited training data, one of the most effective approach is to first pretrain them on large out-of-domain source data and then fine-tune them with the limited target data. The caveat of this is that after fine-tuning the comprehension models tend to perform poorly in the source domain, a phenomenon known as catastrophic forgetting. In this paper, we explore methods that overcome catastrophic forgetting during fine-tuning without assuming access to data from the source domain. We introduce new auxiliary penalty terms and observe the best performance when a combination of auxiliary penalty terms is used to regularise the fine-tuning process for adapting comprehension models. To test our methods, we develop and release 6 narrow domain data sets that could potentially be used as reading comprehension benchmarks.
- 12 November 2020
- Published by Springer Nature
- p. 564-580
Image-Based Table Recognition: Data, Model, and Evaluation
- 12 November 2020
- Published by Springer Nature
- p. 564-580
Image-Based Table Recognition: Data, Model, and Evaluation
- 1 July 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-8
Forget Me Not: Reducing Catastrophic Forgetting for Domain Adaptation in Reading Comprehension
The creation of large-scale open domain reading comprehension data sets in recent years has enabled the development of end-to-end neural comprehension models with promising results. To use these models for domains with limited training data, one of the most effective approach is to first pre-train them on large out-of-domain source data and then fine-tune them with the limited target data. The caveat of this is that after fine-tuning the comprehension models tend to perform poorly in the source domain, a phenomenon known as catastrophic forgetting. In this paper, we explore methods that reduce catastrophic forgetting during fine-tuning without assuming access to data from the source domain. We introduce new auxiliary penalty terms and observe the best performance when a combination of auxiliary penalty terms is used to regularise the fine-tuning process for adapting comprehension models. To test our methods, we develop and release 6 narrow domain data sets that can potentially be used as reading comprehension benchmarks.
- 1 July 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-8
Forget Me Not: Reducing Catastrophic Forgetting for Domain Adaptation in Reading Comprehension
The creation of large-scale open domain reading comprehension data sets in recent years has enabled the development of end-to-end neural comprehension models with promising results. To use these models for domains with limited training data, one of the most effective approach is to first pre-train them on large out-of-domain source data and then fine-tune them with the limited target data. The caveat of this is that after fine-tuning the comprehension models tend to perform poorly in the source domain, a phenomenon known as catastrophic forgetting. In this paper, we explore methods that reduce catastrophic forgetting during fine-tuning without assuming access to data from the source domain. We introduce new auxiliary penalty terms and observe the best performance when a combination of auxiliary penalty terms is used to regularise the fine-tuning process for adapting comprehension models. To test our methods, we develop and release 6 narrow domain data sets that can potentially be used as reading comprehension benchmarks.
- Open Access
Image-based table recognition: data, model, and evaluation
Important information that relates to a specific topic in a document is often organized in tabular format to assist readers with information retrieval and comparison, which may be difficult to provide in natural language. However, tabular data in unstructured digital documents, e.g., Portable Document Format (PDF) and images, are difficult to parse into structured machine-readable format, due to complexity and diversity in their structure and style. To facilitate image-based table recognition with deep learning, we develop the largest publicly available table recognition dataset PubTabNet (https://github.com/ibm-aur-nlp/PubTabNet), containing 568k table images with corresponding structured HTML representation. PubTabNet is automatically generated by matching the XML and PDF representations of the scientific articles in PubMed Central Open Access Subset (PMCOA). We also propose a novel attention-based encoder-dual-decoder (EDD) architecture that converts images of tables into HTML code. The model has a structure decoder which reconstructs the table structure and helps the cell decoder to recognize cell content. In addition, we propose a new Tree-Edit-Distance-based Similarity (TEDS) metric for table recognition, which more appropriately captures multi-hop cell misalignment and OCR errors than the pre-established metric. The experiments demonstrate that the EDD model can accurately recognize complex tables solely relying on the image representation, outperforming the state-of-the-art by 9.7% absolute TEDS score.
Open AccessImage-based table recognition: data, model, and evaluation
Important information that relates to a specific topic in a document is often organized in tabular format to assist readers with information retrieval and comparison, which may be difficult to provide in natural language. However, tabular data in unstructured digital documents, e.g., Portable Document Format (PDF) and images, are difficult to parse into structured machine-readable format, due to complexity and diversity in their structure and style. To facilitate image-based table recognition with deep learning, we develop the largest publicly available table recognition dataset PubTabNet (https://github.com/ibm-aur-nlp/PubTabNet), containing 568k table images with corresponding structured HTML representation. PubTabNet is automatically generated by matching the XML and PDF representations of the scientific articles in PubMed Central Open Access Subset (PMCOA). We also propose a novel attention-based encoder-dual-decoder (EDD) architecture that converts images of tables into HTML code. The model has a structure decoder which reconstructs the table structure and helps the cell decoder to recognize cell content. In addition, we propose a new Tree-Edit-Distance-based Similarity (TEDS) metric for table recognition, which more appropriately captures multi-hop cell misalignment and OCR errors than the pre-established metric. The experiments demonstrate that the EDD model can accurately recognize complex tables solely relying on the image representation, outperforming the state-of-the-art by 9.7% absolute TEDS score.
- Open Access
- 1 January 2020
- Published by Association for Computational Linguistics (ACL)
- p. 195-204
Global Locality in Biomedical Relation and Event Extraction
- 1 January 2020
- Published by Association for Computational Linguistics (ACL)
- p. 195-204
Open AccessGlobal Locality in Biomedical Relation and Event Extraction
- Open Access
- 18 September 2019
- Published by Springer Nature in Scientific Reports
- Vol. 9 (1) , 1-8
Revisiting the use of web search data for stock market movements
Advances in Big Data make it possible to make short-term forecasts for market trends from previously unexplored sources. Trading strategies were recently developed by exploiting a link between the online search activity of certain terms semantically related to finance and market movements. Here we build on these earlier results by exploring a data-driven strategy which adaptively leverages the Google Correlate service and automatically chooses a new set of search terms for every trading decision. In a backtesting experiment run from 2008 to 2017 we obtained a 499% cumulative return which compares favourably with benchmark strategies. A crowdsourcing exercise reveals that the term selection process preferentially selects highly specific terms semantically related to finance (e.g. Wells Fargo Bank), which may capture the transient interests of investors, but at the cost of a shorter span of validity. The adaptive strategy quickly updates the set of search terms when a better combination is found, leading to more consistent predictability. We anticipate that this adaptive decision framework can be of value not only for financial applications, but also in other areas of computational social science, where linkages between facets of collective human behavior and online searches can be inferred from digital footprint data.
- 18 September 2019
- Published by Springer Nature in Scientific Reports
- Vol. 9 (1) , 1-8
Open AccessRevisiting the use of web search data for stock market movements
Advances in Big Data make it possible to make short-term forecasts for market trends from previously unexplored sources. Trading strategies were recently developed by exploiting a link between the online search activity of certain terms semantically related to finance and market movements. Here we build on these earlier results by exploring a data-driven strategy which adaptively leverages the Google Correlate service and automatically chooses a new set of search terms for every trading decision. In a backtesting experiment run from 2008 to 2017 we obtained a 499% cumulative return which compares favourably with benchmark strategies. A crowdsourcing exercise reveals that the term selection process preferentially selects highly specific terms semantically related to finance (e.g. Wells Fargo Bank), which may capture the transient interests of investors, but at the cost of a shorter span of validity. The adaptive strategy quickly updates the set of search terms when a better combination is found, leading to more consistent predictability. We anticipate that this adaptive decision framework can be of value not only for financial applications, but also in other areas of computational social science, where linkages between facets of collective human behavior and online searches can be inferred from digital footprint data.
- Top Cited Papers
- 1 September 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1015-1022
PubLayNet: Largest Dataset Ever for Document Layout Analysis
Recognizing the layout of unstructured digital documents is an important step when parsing the documents into structured machine-readable format for downstream applications. Deep neural networks that are developed for computer vision have been proven to be an effective method to analyze layout of document images. However, document layout datasets that are currently publicly available are several magnitudes smaller than established computing vision datasets. Models have to be trained by transfer learning from a base model that is pre-trained on a traditional computer vision dataset. In this paper, we develop the PubLayNet dataset for document layout analysis by automatically matching the XML representations and the content of over 1 million PDF articles that are publicly available on PubMed Central. The size of the dataset is comparable to established computer vision datasets, containing over 360 thousand document images, where typical document layout elements are annotated. The experiments demonstrate that deep neural networks trained on PubLayNet accurately recognize the layout of scientific articles. The pre-trained models are also a more effective base mode for transfer learning on a different document domain. We release the dataset (https://github.com/ibm-aur-nlp/PubLayNet) to support development and evaluation of more advanced models for document layout analysis.
- 1 September 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1015-1022
Top Cited PapersPubLayNet: Largest Dataset Ever for Document Layout Analysis
Recognizing the layout of unstructured digital documents is an important step when parsing the documents into structured machine-readable format for downstream applications. Deep neural networks that are developed for computer vision have been proven to be an effective method to analyze layout of document images. However, document layout datasets that are currently publicly available are several magnitudes smaller than established computing vision datasets. Models have to be trained by transfer learning from a base model that is pre-trained on a traditional computer vision dataset. In this paper, we develop the PubLayNet dataset for document layout analysis by automatically matching the XML representations and the content of over 1 million PDF articles that are publicly available on PubMed Central. The size of the dataset is comparable to established computer vision datasets, containing over 360 thousand document images, where typical document layout elements are annotated. The experiments demonstrate that deep neural networks trained on PubLayNet accurately recognize the layout of scientific articles. The pre-trained models are also a more effective base mode for transfer learning on a different document domain. We release the dataset (https://github.com/ibm-aur-nlp/PubLayNet) to support development and evaluation of more advanced models for document layout analysis.
- Open Access
PubLayNet: largest dataset ever for document layout analysis
Recognizing the layout of unstructured digital documents is an important step when parsing the documents into structured machine-readable format for downstream applications. Deep neural networks that are developed for computer vision have been proven to be an effective method to analyze layout of document images. However, document layout datasets that are currently publicly available are several magnitudes smaller than established computing vision datasets. Models have to be trained by transfer learning from a base model that is pre-trained on a traditional computer vision dataset. In this paper, we develop the PubLayNet dataset for document layout analysis by automatically matching the XML representations and the content of over 1 million PDF articles that are publicly available on PubMed Central. The size of the dataset is comparable to established computer vision datasets, containing over 360 thousand document images, where typical document layout elements are annotated. The experiments demonstrate that deep neural networks trained on PubLayNet accurately recognize the layout of scientific articles. The pre-trained models are also a more effective base mode for transfer learning on a different document domain. We release the dataset (https://github.com/ibm-aur-nlp/PubLayNet) to support development and evaluation of more advanced models for document layout analysis.
Open AccessPubLayNet: largest dataset ever for document layout analysis
Recognizing the layout of unstructured digital documents is an important step when parsing the documents into structured machine-readable format for downstream applications. Deep neural networks that are developed for computer vision have been proven to be an effective method to analyze layout of document images. However, document layout datasets that are currently publicly available are several magnitudes smaller than established computing vision datasets. Models have to be trained by transfer learning from a base model that is pre-trained on a traditional computer vision dataset. In this paper, we develop the PubLayNet dataset for document layout analysis by automatically matching the XML representations and the content of over 1 million PDF articles that are publicly available on PubMed Central. The size of the dataset is comparable to established computer vision datasets, containing over 360 thousand document images, where typical document layout elements are annotated. The experiments demonstrate that deep neural networks trained on PubLayNet accurately recognize the layout of scientific articles. The pre-trained models are also a more effective base mode for transfer learning on a different document domain. We release the dataset (https://github.com/ibm-aur-nlp/PubLayNet) to support development and evaluation of more advanced models for document layout analysis.