Long Nguyen
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
Long Nguyen has a total of 95 co-authors. Here are the top co-authors by number of shared publications:
- Dien Dinh
University of Science
13 shared publications Dien Dinh
University of Science
11 shared publicationsPhuoc Tran
4 shared publications- Thinh Hung Truong
The University of Melbourne
3 shared publications - Tuan An Dao
The University of Tokyo
3 shared publications - Khang Pham
Faculty of Information Technology, University of Science, Ho Chi Minh, Vietnam and Vietnam National University, Ho Chi Minh, Vietnam
2 shared publications
Latest Publications
- 9 October 2026
- Published by Springer Nature
- p. 123-134
Resource-efficient Knowledge Graph Question Answering with Small Language Models
- 9 October 2026
- Published by Springer Nature
- p. 123-134
Resource-efficient Knowledge Graph Question Answering with Small Language Models
- 2 September 2026
- Published by Springer Nature
- p. 333-345
Optimization Approaches for Language Models in the Task of Translating Sino-Vietnamese Texts into Modern Vietnamese
- 2 September 2026
- Published by Springer Nature
- p. 333-345
Optimization Approaches for Language Models in the Task of Translating Sino-Vietnamese Texts into Modern Vietnamese
- 15 August 2026
- Published by Springer Nature
- p. 405-419
Enhancing RAFT with Knowledge Graphs for Question Answering on Vietnamese Legal Texts
- 15 August 2026
- Published by Springer Nature
- p. 405-419
Enhancing RAFT with Knowledge Graphs for Question Answering on Vietnamese Legal Texts
- Open Access
- 25 April 2026
- Published by Publishing House for Science and Technology, Vietnam Academy of Science and Technology (Publications) in Vietnam Journal of Science and Technology
- Vol. 64 (2)
Adapting knowledge graph embedding for neural machine translation
In the era of deep learning and the rise of Sequence to Sequence architecture, Neural Machine Translation (NMT) has significantly improved in efficiency and performance. However, NMT models still face challenges due to the need for large amounts of training data, particularly for language pairs with insufficient resources, resulting in the corpus sparsity problem. This paper explores the integration of Knowledge Graphs (KGs) into NMT models to enhance the translation of rare and out-of-vocabulary (OOV) words. Specifically, our method, KGE-NMT, leverages structured knowledge from KGs to improve the semantic representation of entities in sentences, thereby enhancing the overall translation quality. Experimental results on English-Vietnamese and English-German language pairs (i.e., IWSLT datasets) show that our KGE-NMT model significantly outperforms baseline models, confirming the benefits of incorporating external knowledge into the machine translation process.
- 25 April 2026
- Published by Publishing House for Science and Technology, Vietnam Academy of Science and Technology (Publications) in Vietnam Journal of Science and Technology
- Vol. 64 (2)
Open AccessAdapting knowledge graph embedding for neural machine translation
In the era of deep learning and the rise of Sequence to Sequence architecture, Neural Machine Translation (NMT) has significantly improved in efficiency and performance. However, NMT models still face challenges due to the need for large amounts of training data, particularly for language pairs with insufficient resources, resulting in the corpus sparsity problem. This paper explores the integration of Knowledge Graphs (KGs) into NMT models to enhance the translation of rare and out-of-vocabulary (OOV) words. Specifically, our method, KGE-NMT, leverages structured knowledge from KGs to improve the semantic representation of entities in sentences, thereby enhancing the overall translation quality. Experimental results on English-Vietnamese and English-German language pairs (i.e., IWSLT datasets) show that our KGE-NMT model significantly outperforms baseline models, confirming the benefits of incorporating external knowledge into the machine translation process.
- 5 April 2026
- Published by Springer Nature
- p. 175-188
ViVG: A Dataset for English–Vietnamese Machine Translation
- 5 April 2026
- Published by Springer Nature
- p. 175-188
ViVG: A Dataset for English–Vietnamese Machine Translation
- 5 April 2026
- Published by Springer Nature
- p. 298-308
MTF-KGC: Multi-Scale Temporal Filtering for Knowledge Graph Completion
- 5 April 2026
- Published by Springer Nature
- p. 298-308
MTF-KGC: Multi-Scale Temporal Filtering for Knowledge Graph Completion
- 5 April 2026
- Published by Springer Nature
- p. 606-620
VKG-IR: Reliability-Aware Visual Graph Embedding for Fine-Grained Image Retrieval
- 5 April 2026
- Published by Springer Nature
- p. 606-620
VKG-IR: Reliability-Aware Visual Graph Embedding for Fine-Grained Image Retrieval
- Open Access
- 28 February 2026
- Published by Ho Chi Minh City University of Technology and Engineering in Journal of Technical Education Science
- Vol. 21 (01) , 99-109
Multilingual Neural Machine Translation for Asian Language Treebank
This study examines multilingual neural machine translation (MNMT) for a diverse group of low-resource Asian languages-Bengali, Filipino, Indonesian, Japanese, Khmer, Malay, and Vietnamese-which differ substantially in linguistic families, writing systems, and typology. This paper evaluates state-of-the-art MNMT systems and introduces a Compact & Language-Sensitive MNMT model designed to improve translation performance while reducing computational cost. The proposed approach shares parameters through a compact multilingual representation, and enhances language discrimination using language-sensitive embeddings, a language-sensitive discriminator, and an adaptive cross-attention mechanism that selects attention parameters based on specific language pairs. Integrated with a multi-stage fine-tuning strategy, this model effectively strengthens cross-lingual transfer while maintaining robust language-specific representations. Experiments on the ALT multi-parallel corpus and the KFTT English-Japanese dataset demonstrate that multilingual models significantly outperform single-language NMT baselines. Despite its smaller size, the proposed Compact & Language-Sensitive MNMT achieves competitive or superior BLEU scores compared to Google’s MNMT, confirming the effectiveness of guided parameter sharing and language-sensitive training. These results highlight the value of compact multilingual architectures and multi-parallel datasets for advancing low-resource Asian machine translation.
- 28 February 2026
- Published by Ho Chi Minh City University of Technology and Engineering in Journal of Technical Education Science
- Vol. 21 (01) , 99-109
Open AccessMultilingual Neural Machine Translation for Asian Language Treebank
This study examines multilingual neural machine translation (MNMT) for a diverse group of low-resource Asian languages-Bengali, Filipino, Indonesian, Japanese, Khmer, Malay, and Vietnamese-which differ substantially in linguistic families, writing systems, and typology. This paper evaluates state-of-the-art MNMT systems and introduces a Compact & Language-Sensitive MNMT model designed to improve translation performance while reducing computational cost. The proposed approach shares parameters through a compact multilingual representation, and enhances language discrimination using language-sensitive embeddings, a language-sensitive discriminator, and an adaptive cross-attention mechanism that selects attention parameters based on specific language pairs. Integrated with a multi-stage fine-tuning strategy, this model effectively strengthens cross-lingual transfer while maintaining robust language-specific representations. Experiments on the ALT multi-parallel corpus and the KFTT English-Japanese dataset demonstrate that multilingual models significantly outperform single-language NMT baselines. Despite its smaller size, the proposed Compact & Language-Sensitive MNMT achieves competitive or superior BLEU scores compared to Google’s MNMT, confirming the effectiveness of guided parameter sharing and language-sensitive training. These results highlight the value of compact multilingual architectures and multi-parallel datasets for advancing low-resource Asian machine translation.
- Open Access
- 1 February 2026
- Published by Wiley in Chemistry–Methods
- Vol. 6 (2)
An In Situ ATR‐FTIR Electrochemical Cell for the Study of Battery Processes: Design, Implementation, and Data Analysis
Phenomena occurring in the electrolyte as well as the interfaces with the electrodes, such as Li + solvation/desolvation and solid–electrolyte interphase formation, govern the performance and safety of Li‐ion batteries. In this article, a dedicated, spring‐loaded operando attenuated‐total‐reflectance Fourier‐transform infrared cell is presented, enabling quantitative, time‐resolved probing of the electrode–electrolyte processes under electrochemical examination. The optical design is based on a 45 ∘ incidence diamond waveguide, while the electrochemical setup comprises a gas‐tight stainless steel body. The procedure for preparing the self‐supported electrode and the acquisition protocol are also presented, together with a reproducible analysis workflow for tracking solvated versus free electrolyte solvent species without baseline subtraction. Representative measurements on composite tin electrodes validate the ability of the setup to resolve band shifts and intensity changes linked to Li + coordination and electrolyte reduction. The methodology generalizes to diverse negative‐electrode chemistries and provides molecular‐level insight into battery phenomena under electrochemical operating conditions.
- 1 February 2026
- Published by Wiley in Chemistry–Methods
- Vol. 6 (2)
Open AccessAn In Situ ATR‐FTIR Electrochemical Cell for the Study of Battery Processes: Design, Implementation, and Data Analysis
Phenomena occurring in the electrolyte as well as the interfaces with the electrodes, such as Li + solvation/desolvation and solid–electrolyte interphase formation, govern the performance and safety of Li‐ion batteries. In this article, a dedicated, spring‐loaded operando attenuated‐total‐reflectance Fourier‐transform infrared cell is presented, enabling quantitative, time‐resolved probing of the electrode–electrolyte processes under electrochemical examination. The optical design is based on a 45 ∘ incidence diamond waveguide, while the electrochemical setup comprises a gas‐tight stainless steel body. The procedure for preparing the self‐supported electrode and the acquisition protocol are also presented, together with a reproducible analysis workflow for tracking solvated versus free electrolyte solvent species without baseline subtraction. Representative measurements on composite tin electrodes validate the ability of the setup to resolve band shifts and intensity changes linked to Li + coordination and electrolyte reduction. The methodology generalizes to diverse negative‐electrode chemistries and provides molecular‐level insight into battery phenomena under electrochemical operating conditions.
- 10 January 2026
- Published by Elsevier in Information Processing & Management
- Vol. 63 (4)
ChartReLA: A compact vision-language model for comprehensive chart reasoning via relationship modeling
- 10 January 2026
- Published by Elsevier in Information Processing & Management
- Vol. 63 (4)
ChartReLA: A compact vision-language model for comprehensive chart reasoning via relationship modeling