Mengtao Sun
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
Mengtao Sun has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
- Ibrahim A. Hameed
Department of Mechanical Engineering and Technology Management, Norwegian University of Life Sciences, Ås, Akershus, 1433, Norway
4 shared publications - Hao Wang
Leiden University
3 shared publications - Mark Pasquine
Department of International Business, Norwegian University of Science and Technology, 6009 Ålesund, Norway
3 shared publications Lei Liu
Beijing Institute for Scientific and Engineering Computing, College of Applied Sciences, Beijing University of Technologyg, Beijin, China
2 shared publications- Carl Petter Skaar Kulseng
Division of Obstetrics and Gynecology Akershus University Hospital Lørenskog Norway
1 shared publication Chunxia Zhang
School of Software, Beijing Institute of Technology, Beijing, China
1 shared publication
Latest Publications
- 17 April 2022
- Published by Springer Nature
- p. 298-308
Skip Truncation for Sentiment Analysis of Long Review Information Based on Grammatical Structures
- 17 April 2022
- Published by Springer Nature
- p. 298-308
Skip Truncation for Sentiment Analysis of Long Review Information Based on Grammatical Structures
- Open Access
- 20 January 2022
- Published by MDPI AG in Information
- Vol. 13 (2) , 49
Learning the Morphological and Syntactic Grammars for Named Entity Recognition
In some languages, Named Entity Recognition (NER) is severely hindered by complex linguistic structures, such as inflection, that will confuse the data-driven models when perceiving the word’s actual meaning. This work tries to alleviate these problems by introducing a novel neural network based on morphological and syntactic grammars. The experiments were performed in four Nordic languages, which have many grammar rules. The model was named the NorG network (Nor: Nordic Languages, G: Grammar). In addition to learning from the text content, the NorG network also learns from the word writing form, the POS tag, and dependency. The proposed neural network consists of a bidirectional Long Short-Term Memory (Bi-LSTM) layer to capture word-level grammars, while a bidirectional Graph Attention (Bi-GAT) layer is used to capture sentence-level grammars. Experimental results from four languages show that the grammar-assisted network significantly improves the results against baselines. We also investigate how the NorG network works on each grammar component by some exploratory experiments.
- 20 January 2022
- Published by MDPI AG in Information
- Vol. 13 (2) , 49
Open AccessLearning the Morphological and Syntactic Grammars for Named Entity Recognition
In some languages, Named Entity Recognition (NER) is severely hindered by complex linguistic structures, such as inflection, that will confuse the data-driven models when perceiving the word’s actual meaning. This work tries to alleviate these problems by introducing a novel neural network based on morphological and syntactic grammars. The experiments were performed in four Nordic languages, which have many grammar rules. The model was named the NorG network (Nor: Nordic Languages, G: Grammar). In addition to learning from the text content, the NorG network also learns from the word writing form, the POS tag, and dependency. The proposed neural network consists of a bidirectional Long Short-Term Memory (Bi-LSTM) layer to capture word-level grammars, while a bidirectional Graph Attention (Bi-GAT) layer is used to capture sentence-level grammars. Experimental results from four languages show that the grammar-assisted network significantly improves the results against baselines. We also investigate how the NorG network works on each grammar component by some exploratory experiments.
- Open Access
- 28 December 2021
- Published by MDPI AG in Applied Sciences
- Vol. 12 (1) , 283
Detecting Small Anatomical Structures in 3D Knee MRI Segmentation by Fully Convolutional Networks
Accurately identifying the pixels of small organs or lesions from magnetic resonance imaging (MRI) has a critical impact on clinical diagnosis. U-net is the most well-known and commonly used neural network for image segmentation. However, the small anatomical structures in medical images cannot be well recognised by U-net. This paper explores the performance of the U-net architectures in knee MRI segmentation to find a relative structure that can obtain high accuracies for both small and large anatomical structures. To maximise the utilities of U-net architecture, we apply three types of components, residual blocks, squeeze-and-excitation (SE) blocks, and dense blocks, to construct four variants of U-net, namely U-net variants. Among these variants, our experiments show that SE blocks can improve the segmentation accuracies of small labels. We adopt DeepLabv3plus architecture for 3D medical image segmentation by equipping SE blocks based on this discovery. The experimental results show that U-net with SE block achieves higher accuracy in parts of small anatomical structures. In contrast, DeepLabv3plus with SE block performs better on the average dice coefficient of small and large labels.
- 28 December 2021
- Published by MDPI AG in Applied Sciences
- Vol. 12 (1) , 283
Open AccessDetecting Small Anatomical Structures in 3D Knee MRI Segmentation by Fully Convolutional Networks
Accurately identifying the pixels of small organs or lesions from magnetic resonance imaging (MRI) has a critical impact on clinical diagnosis. U-net is the most well-known and commonly used neural network for image segmentation. However, the small anatomical structures in medical images cannot be well recognised by U-net. This paper explores the performance of the U-net architectures in knee MRI segmentation to find a relative structure that can obtain high accuracies for both small and large anatomical structures. To maximise the utilities of U-net architecture, we apply three types of components, residual blocks, squeeze-and-excitation (SE) blocks, and dense blocks, to construct four variants of U-net, namely U-net variants. Among these variants, our experiments show that SE blocks can improve the segmentation accuracies of small labels. We adopt DeepLabv3plus architecture for 3D medical image segmentation by equipping SE blocks based on this discovery. The experimental results show that U-net with SE block achieves higher accuracy in parts of small anatomical structures. In contrast, DeepLabv3plus with SE block performs better on the average dice coefficient of small and large labels.
- Open Access
- 17 November 2021
- Published by MDPI AG in Applied Sciences
- Vol. 11 (22) , 10860
Machine Translation in Low-Resource Languages by an Adversarial Neural Network
Existing Sequence-to-Sequence (Seq2Seq) Neural Machine Translation (NMT) shows strong capability with High-Resource Languages (HRLs). However, this approach poses serious challenges when processing Low-Resource Languages (LRLs), because the model expression is limited by the training scale of parallel sentence pairs. This study utilizes adversary and transfer learning techniques to mitigate the lack of sentence pairs in LRL corpora. We propose a new Low resource, Adversarial, Cross-lingual (LAC) model for NMT. In terms of the adversary technique, LAC model consists of a generator and discriminator. The generator is a Seq2Seq model that produces the translations from source to target languages, while the discriminator measures the gap between machine and human translations. In addition, we introduce transfer learning on LAC model to help capture the features in rare resources because some languages share the same subject-verb-object grammatical structure. Rather than using the entire pretrained LAC model, we separately utilize the pretrained generator and discriminator. The pretrained discriminator exhibited better performance in all experiments. Experimental results demonstrate that the LAC model achieves higher Bilingual Evaluation Understudy (BLEU) scores and has good potential to augment LRL translations.
- 17 November 2021
- Published by MDPI AG in Applied Sciences
- Vol. 11 (22) , 10860
Open AccessMachine Translation in Low-Resource Languages by an Adversarial Neural Network
Existing Sequence-to-Sequence (Seq2Seq) Neural Machine Translation (NMT) shows strong capability with High-Resource Languages (HRLs). However, this approach poses serious challenges when processing Low-Resource Languages (LRLs), because the model expression is limited by the training scale of parallel sentence pairs. This study utilizes adversary and transfer learning techniques to mitigate the lack of sentence pairs in LRL corpora. We propose a new Low resource, Adversarial, Cross-lingual (LAC) model for NMT. In terms of the adversary technique, LAC model consists of a generator and discriminator. The generator is a Seq2Seq model that produces the translations from source to target languages, while the discriminator measures the gap between machine and human translations. In addition, we introduce transfer learning on LAC model to help capture the features in rare resources because some languages share the same subject-verb-object grammatical structure. Rather than using the entire pretrained LAC model, we separately utilize the pretrained generator and discriminator. The pretrained discriminator exhibited better performance in all experiments. Experimental results demonstrate that the LAC model achieves higher Bilingual Evaluation Understudy (BLEU) scores and has good potential to augment LRL translations.
- 1 July 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 971-978
An Answer Quality Evaluation Algorithm Based on Semantic in Community-Based Question Answering
CQA(Community-Based Question Answering)is a platform for users' information accessing and knowledge sharing. However, the qualities of answers in CQA are not guaranteed because of redundancy, unreliability, and incompletion. In this paper, an answer quality evaluation algorithm based on semantic is proposed. Seed keywords in a certain field are firstly obtained by topic model and expanded in two ways: context and synonyms, then a CNN(convolutional neural network)model is used to train the quality evaluation model. The experimental result shows that the accuracy of answer quality classification can be effectively improved.
- 1 July 2018
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 971-978
An Answer Quality Evaluation Algorithm Based on Semantic in Community-Based Question Answering
CQA(Community-Based Question Answering)is a platform for users' information accessing and knowledge sharing. However, the qualities of answers in CQA are not guaranteed because of redundancy, unreliability, and incompletion. In this paper, an answer quality evaluation algorithm based on semantic is proposed. Seed keywords in a certain field are firstly obtained by topic model and expanded in two ways: context and synonyms, then a CNN(convolutional neural network)model is used to train the quality evaluation model. The experimental result shows that the accuracy of answer quality classification can be effectively improved.
- 1 July 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 2174-2179
A feature optimization algorithm of concept similarity based on Chinese wikipedia
Concept similarity measure based on feature vector has wide application in various fields, but the problems of polysemy and synonym existing in feature vector affect the similarity measure. We present a feature optimization algorithm based on Chinese Wikipedia which can reduces this effect. First we build a POS feature dictionary (POS-Dic) and a POS Tongyici Cilin(POS-Cilin), and then a new feature vector is used for concept similarity measure. Experiments show that the algorithm effectively reduces the influence of polysemy and synonym on the concept similarity measure.
- 1 July 2017
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 2174-2179
A feature optimization algorithm of concept similarity based on Chinese wikipedia
Concept similarity measure based on feature vector has wide application in various fields, but the problems of polysemy and synonym existing in feature vector affect the similarity measure. We present a feature optimization algorithm based on Chinese Wikipedia which can reduces this effect. First we build a POS feature dictionary (POS-Dic) and a POS Tongyici Cilin(POS-Cilin), and then a new feature vector is used for concept similarity measure. Experiments show that the algorithm effectively reduces the influence of polysemy and synonym on the concept similarity measure.