Publications

2026

1. FA-GPNet: When Gaussian Process Meets Auto-Encoder and FBCSP - A Hybrid Model for Motor Imagery Classification

   Trung M. Pham, Hieu M. Pham, Vi K. Nguyen, and 5 more authors

   In Multi-disciplinary Trends in Artificial Intelligence, 2026

   Abs DOI Bib PDF

   Motor imagery electroencephalography (MI-EEG) decoding is challenged by noisy, non-stationary signals, high variability, and limited labeled data. We propose FA-GPNet, a unified framework that integrates Filter Bank Common Spatial Pattern (FBCSP) for multi-band spectral–spatial feature extraction, a deep autoencoder for nonlinear compression and noise reduction, and a Gaussian Process Classifier (GPC) for probabilistic, uncertainty-aware predictions. Unlike conventional FBCSP pipelines that rely on manual feature selection and deterministic classifiers, FA-GPNet replaces heuristic ranking with data-driven latent representation learning and leverages GP’s Bayesian framework for calibrated outputs. Under within-subject evaluation, FA-GPNet achieves 78.19% mean accuracy on BCI Competition IV-2b, surpassing strong traditional baselines and multiple deep networks, while remaining competitive with CapsNet. On the HCM-IU hand-binary MI dataset, FA-GPNet outperforms the well-optimized classical baseline. These results demonstrate that FA-GPNet offers a robust, reproducible, and efficient solution for MI-EEG decoding.

   @inproceedings{pham_fa-gpnet_2026,
     address = {Singapore},
     title = {{FA}-{GPNet}: {When} {Gaussian} {Process} {Meets} {Auto}-{Encoder} and {FBCSP} - {A} {Hybrid} {Model} for {Motor} {Imagery} {Classification}},
     copyright = {Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License},
     isbn = {9789819549634},
     shorttitle = {{FA}-{GPNet}},
     doi = {10.1007/978-981-95-4963-4_11},
     language = {en},
     booktitle = {Multi-disciplinary {Trends} in {Artificial} {Intelligence}},
     publisher = {Springer Nature},
     author = {Pham, Trung M. and Pham, Hieu M. and Nguyen, Vi K. and Tran, Truong D. and Nguyen, Long S. T. and Nguyen, Duc Q. and Ha, Huong T. T. and Quan, Tho T.},
     editor = {Quan, Thanh Tho and Sombattheera, Chattrakul and Pham, Hoang-Anh and Tran, Ngoc Thinh},
     year = {2026},
     keywords = {Auto-Encoder, Electroencephalography (EEG), Gaussian Process, Hybrid Model, Motor Imagery},
     pages = {127--139},
   }

2025

1. ICLR

   Neural Nonmyopic Bayesian Optimization in Dynamic Cost Settings

   Sang T. Truong, Duc Q. Nguyen, Willie Neiswanger, and 4 more authors

   In Towards Agentic AI for Science: Hypothesis Generation, Comprehension, Quantification, and Validation, Mar 2025

   Abs Bib PDF

   Bayesian optimization (BO) is a popular framework for optimizing black-box functions, leveraging probabilistic models such as Gaussian processes. Conventional BO algorithms, however, assume static query costs, which limit their applicability to real-world problems with dynamic cost structures such as geological surveys or biological sequence design, where query costs vary based on the previous actions. We propose a novel nonmyopic BO algorithm named LookaHES featuring dynamic cost models to address this. LookaHES employs a neural network policy for variational optimization over multi-step lookahead horizons to enable planning under dynamic cost environments. Empirically, we benchmark LookaHES on synthetic functions exhibiting varied dynamic cost structures. We subsequently apply LookaHES to a real-world application in protein sequence design using a large language model policy, demonstrating its scalability and effectiveness in handling multi-step planning in a large and complex query space. LookaHES consistently outperforms its myopic counterparts in synthetic and real-world settings, significantly improving efficiency and solution quality. Our implementation is available at https://github.com/sangttruong/nonmyopia.

   @inproceedings{truong_neural_2025,
     title = {Neural {Nonmyopic} {Bayesian} {Optimization} in {Dynamic} {Cost} {Settings}},
     copyright = {Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License},
     url = {https://openreview.net/forum?id=DFCT9dHkko},
     language = {en},
     urldate = {2025-09-08},
     booktitle = {Towards {Agentic} {AI} for {Science}: {Hypothesis} {Generation}, {Comprehension}, {Quantification}, and {Validation}},
     author = {Truong, Sang T. and Nguyen, Duc Q. and Neiswanger, Willie and Griffiths, Ryan-Rhys and Ermon, Stefano and Haber, Nick and Koyejo, Sanmi},
     month = mar,
     year = {2025},
   }

2. AIED

   Riding on the Back of a Whale: A Hackathon Framework for Introducing High School Students to Large Language Models

   Duc Nguyen, Dong Le, Long Nguyen, and 11 more authors

   In Artificial Intelligence in Education. Posters and Late Breaking Results, Workshops and Tutorials, Industry and Innovation Tracks, Practitioners, Doctoral Consortium, Blue Sky, and WideAIED, Mar 2025

   ISSN: 1865-0937

   Abs DOI Bib PDF

   As large language models (LLMs) become more integrated into daily life, it is crucial to foster AI literacy among high school students. However, most AI courses target college-level learners and assume prior knowledge, while high schools often lack the foundational...

   @inproceedings{nguyen_riding_2025,
     title = {Riding on the {Back} of a {Whale}: {A} {Hackathon} {Framework} for {Introducing} {High} {School} {Students} to {Large} {Language} {Models}},
     copyright = {Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License},
     isbn = {978-3-031-99264-3},
     shorttitle = {Riding on the {Back} of a {Whale}},
     url = {https://link.springer.com/chapter/10.1007/978-3-031-99264-3_25},
     doi = {10.1007/978-3-031-99264-3_25},
     language = {en},
     urldate = {2025-09-08},
     booktitle = {Artificial {Intelligence} in {Education}. {Posters} and {Late} {Breaking} {Results}, {Workshops} and {Tutorials}, {Industry} and {Innovation} {Tracks}, {Practitioners}, {Doctoral} {Consortium}, {Blue} {Sky}, and {WideAIED}},
     publisher = {Springer, Cham},
     author = {Nguyen, Duc and Le, Dong and Nguyen, Long and Vu, Quyen and Le, Tran and Nguyen, Dung and Huynh, Nga and Nguyen, Huong and Tran, Phat and Le, Dang and Truong, Sang and Koyejo, Sanmi and Le, Cuong and Quan, Tho},
     year = {2025},
     note = {ISSN: 1865-0937},
     pages = {201--209},
   }

3. EMNLP

   The Sound of Syntax: Finetuning and Comprehensive Evaluation of Language Models for Speech Pathology

   Fagun Patel, Duc Q. Nguyen, Sang T. Truong, and 3 more authors

   In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, Oct 2025

   Abs Bib PDF

   According to the U.S. National Institutes of Health, more than 3.4 million children experience speech disorders that require clinical intervention. The number of speech-language pathologists (SLPs) is roughly 20 times fewer than the number of affected children, highlighting a significant gap in children’s care and a pressing need for technological support that improves the productivity of SLPs. State-of-the-art multimodal language models (MLMs) show promise for supporting SLPs, but their use remains underexplored largely due to a limited understanding of their performance in high-stakes clinical settings. To address this gap, we collaborate with domain experts to develop a taxonomy of real-world use cases of MLMs in speech-language pathologies. Building on this taxonomy, we introduce the first comprehensive benchmark for evaluating MLM across five core use cases, each containing 1,000 manually annotated data points. This benchmark includes robustness and sensitivity tests under various settings, including background noise, speaker gender, and accent. Our evaluation of 15 state-of-the-art MLMs reveals that no single model consistently outperforms others across all tasks. Notably, we find systematic disparities, with models performing better on male speakers, and observe that chain-of-thought prompting can degrade performance on classification tasks with large label spaces and narrow decision boundaries. Furthermore, we study fine-tuning MLMs on domain-specific data, achieving improvements of over 30% compared to base models. These findings highlight both the potential and limitations of current MLMs for speech-language pathology applications, underscoring the need for further research and targeted development.

   @inproceedings{patel_sound_2025,
     address = {Suzhou, China},
     title = {The {Sound} of {Syntax}: {Finetuning} and {Comprehensive} {Evaluation} of {Language} {Models} for {Speech} {Pathology}},
     copyright = {Creative Commons Attribution-ShareAlike 4.0 International License},
     isbn = {979-8-89176-332-6},
     shorttitle = {The {Sound} of {Syntax}},
     url = {https://aclanthology.org/2025.emnlp-main.1768/},
     urldate = {2025-11-05},
     booktitle = {Proceedings of the 2025 {Conference} on {Empirical} {Methods} in {Natural} {Language} {Processing}},
     publisher = {Association for Computational Linguistics},
     author = {Patel, Fagun and Nguyen, Duc Q. and Truong, Sang T. and Vaynshtok, Jody and Koyejo, Sanmi and Haber, Nick},
     editor = {Christodoulopoulos, Christos and Chakraborty, Tanmoy and Rose, Carolyn and Peng, Violet},
     month = oct,
     year = {2025},
     pages = {34895--34913},
   }

4. IJCNN

   Bridging LLMs and Symbolic Reasoning in Educational QA Systems: Insights from the XAI Challenge at IJCNN 2025

   Long S. T. Nguyen, Khang H. N. Vo, Thu H. A. Nguyen, and 13 more authors

   Aug 2025

   arXiv:2508.01263 [cs]

   Abs DOI Bib PDF

   The growing integration of Artificial Intelligence (AI) into education has intensified the need for transparency and interpretability. While hackathons have long served as agile environments for rapid AI prototyping, few have directly addressed eXplainable AI (XAI) in real-world educational contexts. This paper presents a comprehensive analysis of the XAI Challenge 2025, a hackathon-style competition jointly organized by Ho Chi Minh City University of Technology (HCMUT) and the International Workshop on Trustworthiness and Reliability in Neurosymbolic AI (TRNS-AI), held as part of the International Joint Conference on Neural Networks (IJCNN 2025). The challenge tasked participants with building Question-Answering (QA) systems capable of answering student queries about university policies while generating clear, logic-based natural language explanations. To promote transparency and trustworthiness, solutions were required to use lightweight Large Language Models (LLMs) or hybrid LLM-symbolic systems. A high-quality dataset was provided, constructed via logic-based templates with Z3 validation and refined through expert student review to ensure alignment with real-world academic scenarios. We describe the challenge’s motivation, structure, dataset construction, and evaluation protocol. Situating the competition within the broader evolution of AI hackathons, we argue that it represents a novel effort to bridge LLMs and symbolic reasoning in service of explainability. Our findings offer actionable insights for future XAI-centered educational systems and competitive research initiatives.

   @misc{nguyen_bridging_2025,
     title = {Bridging {LLMs} and {Symbolic} {Reasoning} in {Educational} {QA} {Systems}: {Insights} from the {XAI} {Challenge} at {IJCNN} 2025},
     copyright = {Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License},
     shorttitle = {Bridging {LLMs} and {Symbolic} {Reasoning} in {Educational} {QA} {Systems}},
     url = {http://arxiv.org/abs/2508.01263},
     doi = {10.48550/arXiv.2508.01263},
     urldate = {2026-01-01},
     publisher = {arXiv},
     author = {Nguyen, Long S. T. and Vo, Khang H. N. and Nguyen, Thu H. A. and Bui, Tuan C. and Nguyen, Duc Q. and Tran, Thanh-Tung and Nguyen, Anh D. and Nguyen, Minh L. and Baldacci, Fabien and Bui, Thang H. and Nardo, Emanuel Di and Ciaramella, Angelo and Le, Son H. and Ullah, Ihsan and Rocco, Lorenzo Di and Quan, Tho T.},
     month = aug,
     year = {2025},
     note = {arXiv:2508.01263 [cs]},
     keywords = {Computer Science - Artificial Intelligence, Computer Science - Computation and Language, Computer Science - Logic in Computer Science},
     annote = {Comment: The XAI Challenge @ TRNS-AI Workshop, IJCNN 2025: Explainable AI for Educational Question Answering. Website: https://sites.google.com/view/trns-ai/challenge/},
   }

2024

1. AAAI

   Revitalizing Bahnaric Language through Neural Machine Translation: Challenges, Strategies, and Promising Outcomes

   Hoang Nhat Khang Vo, Duc Dong Le, Tran Minh Dat Phan, and 6 more authors

   In Proceedings of the AAAI Conference on Artificial Intelligence, Mar 2024

   Abs DOI Bib PDF

   The Bahnar, a minority ethnic group in Vietnam with ancient roots, hold a language of deep cultural and historical significance. The government is prioritizing the preservation and dissemination of Bahnar language through online availability and cross-generational communication. Recent AI advances, including Neural Machine Translation (NMT), have transformed translation with improved accuracy and fluency, fostering language revitalization through learning, communication, and documentation. In particular, NMT enhances accessibility for Bahnar language speakers, making information and content more available. However, translating Vietnamese to Bahnar language faces practical hurdles due to resource limitations, particularly in the case of Bahnar language as an extremely low-resource language. These challenges encompass data scarcity, vocabulary constraints, and a lack of fine-tuning data. To address these, we propose transfer learning from selected pre-trained models to optimize translation quality and computational efficiency, capitalizing on linguistic similarities between Vietnamese and Bahnar language. Concurrently, we apply tailored augmentation strategies to adapt machine translation for the Vietnamese-Bahnar language context. Our approach is validated through superior results on bilingual Vietnamese-Bahnar language datasets when compared to baseline models. By tackling translation challenges, we help revitalize Bahnar language, ensuring information flows freely and the language thrives.

   @inproceedings{vo_revitalizing_2024,
     title = {Revitalizing {Bahnaric} {Language} through {Neural} {Machine} {Translation}: {Challenges}, {Strategies}, and {Promising} {Outcomes}},
     volume = {38},
     copyright = {Copyright (c) 2024 Association for the Advancement of Artificial Intelligence},
     shorttitle = {Revitalizing {Bahnaric} {Language} through {Neural} {Machine} {Translation}},
     url = {https://ojs.aaai.org/index.php/AAAI/article/view/30385},
     doi = {10.1609/aaai.v38i21.30385},
     language = {en},
     urldate = {2025-09-08},
     booktitle = {Proceedings of the {AAAI} {Conference} on {Artificial} {Intelligence}},
     author = {Vo, Hoang Nhat Khang and Le, Duc Dong and Phan, Tran Minh Dat and Nguyen, Tan Sang and Pham, Quoc Nguyen and Tran, Ngoc Oanh and Nguyen, Duc Q. and Vo, Tran Minh Hieu and Quan, Tho},
     month = mar,
     year = {2024},
     keywords = {Translation-oriented Bahnaric Augmentation},
     pages = {23360--23368},
   }

2. ESWA

   Supervised learning models for social bot detection: Literature review and benchmark

   Hoang-Dung Nguyen, Duc Q. Nguyen, Cong-Duy Nguyen, and 8 more authors

   Expert Systems with Applications, Mar 2024

   Abs DOI Bib PDF

   Over the past decades, online social networks such as Twitter and Facebook have become a significant part of people’s daily lives, particularly amid the ongoing global calamity — the COVID-19 pandemic. This gives room for social bot attacks that are designed to automatically replicate the behavior of real accounts. Most of these bots are employed for nefarious purposes such as disseminating false information, artificially amplifying the popularity of a person or movement, or spreading spam. Many studies have been conducted in an attempt to discover new strategies for identifying social bot accounts. To deal with large-scale attacks from social bots, Machine Learning has emerged as a noticeable path of bot detection problem that can handle massive amounts of data. However, the heterogeneity between studies in terms of problem statements, proposed processes, datasets, and evaluation metrics makes it difficult to assess and compare the efficiency of proposed methods. In this paper, we propose a systematic view of supervised learning methodologies for tweet-based social bot detection, ranging from shallow learning to specific deep learning models. In addition, we introduce a framework that measures various performance aspects and summarizes the in-depth analysis of the results, which were obtained using two datasets comprising 26224 labeled Twitter accounts. The results of this framework, we believe, will be beneficial as a practical guideline for other bot detection research or applications that require the use of machine learning techniques.

   @article{nguyen_supervised_2024,
     title = {Supervised learning models for social bot detection: {Literature} review and benchmark},
     volume = {238},
     copyright = {Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License},
     issn = {0957-4174},
     shorttitle = {Supervised learning models for social bot detection},
     url = {https://www.sciencedirect.com/science/article/pii/S0957417423027197},
     doi = {10.1016/j.eswa.2023.122217},
     urldate = {2025-09-08},
     journal = {Expert Systems with Applications},
     author = {Nguyen, Hoang-Dung and Nguyen, Duc Q. and Nguyen, Cong-Duy and To, Phong T. and Nguyen, Danh H. and Nguyen-Gia, Huy and Tran, Long H. and Tran, Anh Q. and Dang-Hieu, An and Nguyen-Duc, Anh and Quan, Tho},
     month = mar,
     year = {2024},
     keywords = {Machine learning, Benchmark, Social bot detection},
     pages = {122217},
   }

3. ICLR

   Hybrid Transformer and Holt-Winter’s Method for Time Series Forecasting

   Nhi Ngoc Truong, Duc Q. Nguyen, Jeffrey Gropp, and 1 more author

   In ICLR 2024 Workshop on Learning from Time Series For Health, Mar 2024

   Abs Bib PDF

   Time series forecasting is an important research topic in machine learning due to its prevalence in social and scientific applications. Multi-model forecasting paradigm, including model hybridization and model combination, is shown to be more effective than single-model forecasting in the M4 competition. In this study, we hybridize exponential smoothing with transformer architecture to capture both levels and seasonal patterns while exploiting the complex non-linear trend in time series data. We show that our model can capture complex trends and seasonal patterns with moderately improvement in comparison to the state-of-the-arts result from the M4 competition.

   @inproceedings{truong_hybrid_2024,
     title = {Hybrid {Transformer} and {Holt}-{Winter}'s {Method} for {Time} {Series} {Forecasting}},
     copyright = {Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License},
     url = {https://openreview.net/forum?id=LDJUIeb9ut},
     language = {en},
     urldate = {2025-09-08},
     booktitle = {{ICLR} 2024 {Workshop} on {Learning} from {Time} {Series} {For} {Health}},
     author = {Truong, Nhi Ngoc and Nguyen, Duc Q. and Gropp, Jeffrey and Truong, Sang T.},
     month = mar,
     year = {2024},
   }

4. NAACL

   Crossing Linguistic Horizons: Finetuning and Comprehensive Evaluation of Vietnamese Large Language Models

   Sang Truong, Duc Nguyen, Toan Nguyen, and 4 more authors

   In Findings of the Association for Computational Linguistics: NAACL 2024, Jun 2024

   Abs DOI Bib PDF

   Recent advancements in large language models (LLMs) have underscored their importance in the evolution of artificial intelligence. However, despite extensive pretraining on multilingual datasets, available open-sourced LLMs exhibit limited effectiveness in processing Vietnamese. The challenge is exacerbated by the absence of systematic benchmark datasets and metrics tailored for Vietnamese LLM evaluation. To mitigate these issues, we have finetuned LLMs specifically for Vietnamese and developed a comprehensive evaluation framework encompassing 10 tasks and 31 metrics. We observe that finetuning can help LLMs transfer knowledge across languages, serving as an efficient way to bolster their capabilities in non-English languages. Moreover, our analysis indicates that larger models can introduce more biases and uncalibrated outputs and the key factor influencing LLM performance is the quality of the training or finetuning datasets. These insights underscore the significance of meticulous finetuning with high-quality datasets in enhancing LLM performance.

   @inproceedings{truong_crossing_2024,
     address = {Mexico City, Mexico},
     title = {Crossing {Linguistic} {Horizons}: {Finetuning} and {Comprehensive} {Evaluation} of {Vietnamese} {Large} {Language} {Models}},
     copyright = {Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License},
     shorttitle = {Crossing {Linguistic} {Horizons}},
     url = {https://aclanthology.org/2024.findings-naacl.182/},
     doi = {10.18653/v1/2024.findings-naacl.182},
     urldate = {2025-09-08},
     booktitle = {Findings of the {Association} for {Computational} {Linguistics}: {NAACL} 2024},
     publisher = {Association for Computational Linguistics},
     author = {Truong, Sang and Nguyen, Duc and Nguyen, Toan and Le, Dong and Truong, Nhi and Quan, Tho and Koyejo, Sanmi},
     editor = {Duh, Kevin and Gomez, Helena and Bethard, Steven},
     month = jun,
     year = {2024},
     pages = {2849--2900},
   }

5. IEEE Access

   Explainable Neural Subgraph Matching With Learnable Multi-Hop Attention

   Duc Q. Nguyen, Thanh Toan Nguyen, Jun Jo, and 3 more authors

   IEEE Access, Jun 2024

   Abs DOI Bib PDF

   Subgraph matching is a challenging problem with a wide range of applications in drug discovery, social network analysis, biochemistry, and cognitive science. It involves determining whether a given query graph is present within a larger target graph. Traditional graph-matching algorithms provide precise results but face challenges in large graph instances due to the NP-complete nature of the problem, limiting their practical applicability. In contrast, recent neural network-based approximations offer more scalable solutions but often lack interpretable node correspondences. To address these limitations, this article presents a multi-task learning framework called xNeuSM: Explainable Neural Subgraph Matching, which introduces Graph Learnable Multi-hop Attention Networks (GLeMA) that adaptively learn the parameters governing the attention factor decay for each node across hops rather than relying on fixed hyperparameters. Our framework jointly optimizes both subgraph matching and finding subgraph-isomorphism mappings. We provide a theoretical analysis establishing error bounds for GLeMA’s approximation of multi-hop attention as a function of the number of hops. Additionally, we prove that learning distinct attention decay factors for each node leads to a correct approximation of multi-hop attention. Empirical evaluation on real-world datasets shows that xNeuSM achieves substantial improvements in prediction F1 score of up to 34% compared to approximate baselines and, notably, at least a seven-fold faster query time than exact algorithms. With these results, xNeuSM can be applied to solve matching problems in various domains spanning from biochemistry to social science.

   @article{nguyen_explainable_2024,
     title = {Explainable {Neural} {Subgraph} {Matching} {With} {Learnable} {Multi}-{Hop} {Attention}},
     volume = {12},
     copyright = {Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License},
     issn = {2169-3536},
     url = {https://ieeexplore.ieee.org/document/10677500},
     doi = {10.1109/ACCESS.2024.3458050},
     urldate = {2025-09-08},
     journal = {IEEE Access},
     author = {Nguyen, Duc Q. and Toan Nguyen, Thanh and Jo, Jun and Poux, Florent and Anirban, Shikha and Quan, Tho T.},
     year = {2024},
     keywords = {graph neural networks, Graph neural networks, Spread spectrum communication, Pattern matching, subgraph matching, Adaptation models, Explainability, learnable multi-hop attention, Multitasking, Object recognition},
     pages = {130474--130492},
   }

2023

1. IJCNN

   10X Faster Subgraph Matching: Dual Matching Networks with Interleaved Diffusion Attention

   Thanh Toan Nguyen, Duc Q. Nguyen, Zhao Ren, and 3 more authors

   In 2023 International Joint Conference on Neural Networks (IJCNN), Jun 2023

   ISSN: 2161-4407

   Abs DOI Bib PDF

   The goal of subgraph matching is to determine the presence of a particular query pattern within a large collection of data graphs. Despite being a hard problem, subgraph matching is essential in various disciplines, including bioinformatics, text matching, and graph retrieval. Although traditional approaches could provide exact solutions, their computations are known to be NP-complete, leading to an overwhelmingly querying latency. While recent neural-based approaches have been shown to improve the response time, the oversimplified assumption of the first-order network may neglect the generalisability of fully capturing patterns in varying sizes, causing the performance to drop significantly in datasets in various domains. To overcome these limitations, this paper proposes xDualSM, a dual matching neural network model with interleaved diffusion attention. Specifically, we first embed the structural information of graphs into different adjacency matrices, which explicitly capture the intra-graph and cross-graph structures between the query pattern and the target graph. Then, we introduce a dual matching network with interleaved diffusion attention to carefully capture intra-graph and cross-graph information while reducing computational complexity. Empirically, our proposed framework not only boosted the speed of subgraph matching more than 10x compared to the fastest baseline but also achieved significant improvements of 47.64% in Recall and 34.39% in F1-score compared to the state-of-the-art approximation approach on COX2 dataset. In addition, our results are comparable with exact methods.

   @inproceedings{nguyen_10x_2023,
     title = {{10X} {Faster} {Subgraph} {Matching}: {Dual} {Matching} {Networks} with {Interleaved} {Diffusion} {Attention}},
     copyright = {Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License},
     shorttitle = {{10X} {Faster} {Subgraph} {Matching}},
     url = {https://ieeexplore.ieee.org/document/10191159},
     doi = {10.1109/IJCNN54540.2023.10191159},
     urldate = {2025-09-08},
     booktitle = {2023 {International} {Joint} {Conference} on {Neural} {Networks} ({IJCNN})},
     author = {Nguyen, Thanh Toan and Nguyen, Duc Q. and Ren, Zhao and Jo, Jun and Nguyen, Quoc Viet Hung and Nguyen, Thanh Tam},
     month = jun,
     year = {2023},
     note = {ISSN: 2161-4407},
     keywords = {Bioinformatics, graph neural networks, Computational complexity, diffusion attention, graph mining, Impedance matching, matching by embedding, Neural networks, Pattern matching, subgraph matching, Time factors},
     pages = {1--9},
   }

2. ICML

   Thomas: Learning to Explore Human Preference via Probabilistic Reward Model

   Sang T. Truong, Duc Q. Nguyen, Tho Quan, and 1 more author

   In ICML 2023 Workshop The Many Facets of Preference-Based Learning, Oct 2023

   Abs Bib PDF

   Recent breakthroughs in large language models and multimodal models underscore the impressive strides deep learning has made in tackling sophisticated tasks previously deemed achievable solely by humans. In particular, discerning human thoughts or interests via communication and feedback is garnering attention for its potential to enable machines to provide insightful responses or recommendations. Nonetheless, despite progressive developments, preference learning from human feedback is hindered by poor sample complexity, as it primarily employs preferred responses for tuning, consequently failing to holistically capture user preferences. Moreover, it is imperative to ensure diversity in the responses generated, as this diversity is instrumental in enabling users to ascertain their genuine preferences, which in turn, is conducive to the fine-tuning of the response generation model. In this study, we introduce a novel method known as Thomas, which utilizes Bayesian neural networks for capturing user preferences, and Thompson sampling to enhance the exploration ability of the response generation model. This synergy ensures alignment of generated responses with user preferences, while preserving diversity, thus expediting the learning process. Experimental evaluations in synthetic environments affirm the proficiency of our method in swiftly adapting to user preferences and generating increasingly favored responses.

   @inproceedings{truong_thomas_2023,
     title = {Thomas: {Learning} to {Explore} {Human} {Preference} via {Probabilistic} {Reward} {Model}},
     copyright = {Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License},
     shorttitle = {Thomas},
     url = {https://openreview.net/forum?id=LXbGz5aSDo#all},
     language = {en},
     urldate = {2025-09-08},
     booktitle = {{ICML} 2023 {Workshop} {The} {Many} {Facets} of {Preference}-{Based} {Learning}},
     author = {Truong, Sang T. and Nguyen, Duc Q. and Quan, Tho and Koyejo, Sanmi},
     month = oct,
     year = {2023},
   }

3. Low-Rank Adaptation Approach for Vietnamese-Bahnaric Lexical Mapping from Non-Parallel Corpora

   Huy Cam La, Minh Quang Le, Oanh Ngoc Tran, and 5 more authors

   VNUHCM Journal of Engineering and Technology, Oct 2023

   Abs DOI Bib PDF

   Bilingual dictionaries are vital tools for automated machine translation. Leveraging advanced machine learning techniques, it is possible to construct bilingual dictionaries by automatically learning lexical mappings from bilingual corpora. However, procuring extensive bilingual corpora for low-resource languages, such as Bahnaric, poses a significant challenge. Recent studies suggest that non-parallel corpora, supplemented with a handful of anchor words, can aid in the learning of these mappings, which contain parameters for automated translation between source and target languages. The prevailing methodology involves using Generative Adversarial Networks (GANs) and solving the Procrustes orthogonal problem to generate this mapping. This approach, while innovative, exhibits instability and demands substantial computational resources, posing potential issues in rural regions where Bahnaric is spoken natively. To mitigate this, we propose a low-rank adaptation strategy, where the limitations of GANs can be circumvented by directly calculating the rigid transformation between the source and target languages. We evaluated our approach using the French-English dataset, and a low-resource dataset, Vietnamese-Bahnaric. Notably, the Vietnamese-Bahnaric lexical mapping produced by our method is valuable not only to the field of computer science, but also contributes significantly to the preservation of Bahnaric cultural heritage within Vietnam’s ethnic minority communities.

   @article{la_low-rank_2023,
     title = {Low-{Rank} {Adaptation} {Approach} for {Vietnamese}-{Bahnaric} {Lexical} {Mapping} from {Non}-{Parallel} {Corpora}},
     volume = {6},
     copyright = {Copyright (c) 2024 Huy Cam La, Minh Quang Le, Oanh Ngoc Tran, Dong Duc Le, Duc Q. Nguyen, Sang Tan Nguyen, Quan Tran, Tho Thanh Quan},
     issn = {2615-9872},
     url = {https://stdjet.scienceandtechnology.com.vn/index.php/stdjet/article/view/1197},
     doi = {10.32508/stdjet.v6iSI8.1197},
     language = {en},
     number = {SI8},
     urldate = {2025-09-08},
     journal = {VNUHCM Journal of Engineering and Technology},
     author = {La, Huy Cam and Le, Minh Quang and Tran, Oanh Ngoc and Le, Dong Duc and Nguyen, Duc Q. and Nguyen, Sang Tan and Tran, Quan and Quan, Tho Thanh},
     year = {2023},
     keywords = {Kabsch algorithm, lexical mapping, Low-rank adaptation, low-resource language},
     pages = {19--32},
   }

4. Voice conversion for natural-Sounding speech generation on low-Resource languages: A case study of bahnaric

   Dat T. Dang, Thai Q. Tang, Duc Q. Nguyen, and 2 more authors

   VNUHCM Journal of Engineering and Technology, Oct 2023

   Abs DOI Bib PDF

   Bahnar is an ethnic minority group in Vietnam, prioritized by the government for the preservation of their cultural heritage, traditions, and language. In the current era of AI technology, there is substantial potential in synthesizing Bahnar voices to support these preservation endeavors. While voice conversion technology has made strides in enhancing the quality and naturalness of synthesized speech, its focus has predominantly been on widely spoken languages. Consequently, low-resource languages like the Bahnaric language family encounter numerous disadvantages in voice synthesis. This study addresses the formidable challenge of synthesizing natural-sounding speech in low-resource languages by exploring the application of voice conversion techniques to the Bahnaric language. We introduce the BN-TTS-VC system, a pioneering approach that integrates a text-to-speech system based on Grad-TTS with voice conversion techniques derived from StarGANv2-VC, both tailored specifically for the nuances of the Bahnaric language. Grad-TTS allows the system to articulate Bahnaric words without vocabulary limitations, while StarGANv2-VC enhances the naturalness of synthesized speech, particularly in the context of low-resource languages like Bahnaric. Moreover, we introduce the Bahnaric-fine-tuned HiFi-GAN model to further enhance voice quality with native accents, ensuring a more authentic representation of Bahnaric speech. To assess the effectiveness of our approach, we conducted experiments based on human evaluations from volunteers. The preliminary results are promising, indicating the potential of our methodology in synthesizing natural-sounding Bahnaric speech. Through this research, we aim to make significant contributions to the ongoing efforts to preserve and promote the linguistic and cultural heritage of the Bahnar ethnic minority group. By leveraging the power of AI technology, we aspire to bridge the gap in speech synthesis for low-resource languages and facilitate the preservation of their invaluable cultural heritage.

   @article{dat_voice_2023,
     title = {Voice conversion for natural-{Sounding} speech generation on low-{Resource} languages: {A} case study of bahnaric},
     volume = {6},
     copyright = {Copyright (c) 2024 Dang Tran Dat, Tang Quoc Thai, Nguyen Quang Duc, Vo Duy Hung, Quan Thanh Tho},
     issn = {2615-9872},
     shorttitle = {Voice conversion for natural-{Sounding} speech generation on low-{Resource} languages},
     url = {https://stdjet.scienceandtechnology.com.vn/index.php/stdjet/article/view/1198},
     doi = {10.32508/stdjet.v6iSI8.1198},
     language = {en},
     number = {SI8},
     urldate = {2025-09-08},
     journal = {VNUHCM Journal of Engineering and Technology},
     author = {Dang, Dat T. and Tang, Thai Q. and Nguyen, Duc Q. and Vo, Hung D. and Quan, Tho T.},
     year = {2023},
     pages = {33--45},
   }

5. A Novel Approach for Non-Query Fake News Detection Using K-SOM and Graph Neural Networks

   Hoang-Danh Nguyen, Thanh-Phong To, Duc Q. Nguyen, and 4 more authors

   In 2023 15th International Conference on Knowledge and Systems Engineering (KSE), Oct 2023

   ISSN: 2694-4804

   Abs DOI Bib PDF

   The widespread use of social networks has led to an increase in the number of users and posts on these platforms. However, the proliferation of fake news, particularly in the healthcare sector due to the COVID-19 pandemic, has become a significant concern. This has become a significant concern as accepting such fake news can have severe consequences on the health and lives of those who are exposed to it, leading to confusion, social disorder, and reputational damage for individuals, organizations, and businesses. Therefore, the automatic detection of fake news on social networks is of utmost importance. In this study, we propose the FANSOG model, which utilizes K-SOM to cluster articles and a graph-based model to automatically detect non-query fake news. Our findings demonstrate that the FANSOG model outperforms other state-of-the-art models in the same research direction.

   @inproceedings{nguyen_novel_2023,
     title = {A {Novel} {Approach} for {Non}-{Query} {Fake} {News} {Detection} {Using} {K}-{SOM} and {Graph} {Neural} {Networks}},
     copyright = {Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License},
     url = {https://ieeexplore.ieee.org/document/10299416},
     doi = {10.1109/KSE59128.2023.10299416},
     urldate = {2025-09-08},
     booktitle = {2023 15th {International} {Conference} on {Knowledge} and {Systems} {Engineering} ({KSE})},
     author = {Nguyen, Hoang-Danh and To, Thanh-Phong and Nguyen, Duc Q. and Huynh, Thanh-Trung and Tran, Tham and Bui, Cong-Tuan and Quan, Tho},
     month = oct,
     year = {2023},
     note = {ISSN: 2694-4804},
     keywords = {Social networking (online), Graph neural networks, Analytical models, Fake News Detection, Graph Neural Network, K-SOM clustering, Pandemics, Predictive models, Text categorization, Training},
     pages = {1--6},
   }

6. Unlocking the Potential: an evaluation of Text-to-Speech Models for the Bahnar Language

   Giang L. Dinh, Tho T. Quan, Duc Q. Nguyen, and 2 more authors

   In 2023 IEEE/ACIS 8th International Conference on Big Data, Cloud Computing, and Data Science (BCD), Oct 2023

   ISSN: 2835-4419

   Abs DOI Bib PDF

   The paper aims at evaluating the effectiveness of an AI based mobile application of text- to-speech models for Bahnar language. In this application, a sequential combination of two models was implemented, starting with the application of the Grad-TTS model and subsequently followed by the Hifi-GAN model. Grad-TTS was employed to ensure a highly correct pronunciation of Bahnar words without being constrained by the dataset. The strengths of Hifi-GAN, in other hands, have been fine-tuned for the Bahnaric language to enhance the quality of synthesized audio, inorder to produce a native-like Bahnar voice and accent. Those artificially generated sounds from our model achieved a high level of naturalness.

   @inproceedings{lu_unlocking_2023,
     title = {Unlocking the {Potential}: an evaluation of {Text}-to-{Speech} {Models} for the {Bahnar} {Language}},
     copyright = {Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License},
     shorttitle = {Unlocking the {Potential}},
     url = {https://ieeexplore.ieee.org/document/10466279},
     doi = {10.1109/BCD57833.2023.10466279},
     urldate = {2025-09-08},
     booktitle = {2023 {IEEE}/{ACIS} 8th {International} {Conference} on {Big} {Data}, {Cloud} {Computing}, and {Data} {Science} ({BCD})},
     author = {Dinh, Giang L. and Quan, Tho T. and Nguyen, Duc Q. and Vu, Hai H. and Nguyen, Quy T.},
     month = oct,
     year = {2023},
     note = {ISSN: 2835-4419},
     keywords = {Bahnar language, Big Data, Cloud computing, Computational modeling, Data models, Data science, Mean Opinion Score (MOS), modified rhyme test, Natural languages, speech synthesis, text-to-speech conversion, Vocoders},
     pages = {126--129},
   }

7. BaNaVA: A cross-platform AI mobile application for preserving the Bahnaric languages

   Tho T. Quan, Giang L. Dinh, Duc Q. Nguyen, and 4 more authors

   In 2023 IEEE/ACIS 8th International Conference on Big Data, Cloud Computing, and Data Science (BCD), Oct 2023

   ISSN: 2835-4419

   Abs DOI Bib PDF

   AI-powered translation is a promising solution to the language barrier faced by the Bahnar people. However, developing low-resource text-to-speech translation systems is challenging. The authors propose a mobile application called BaNaVA to address these challenges. BaNaVA uses a combination of natural machine translation and linguistic analysis to translate between Vietnamese and Bahnaric with high accuracy, while also using a voice conversion system to convert the voice quality to match that of a genuine Bahnar individual. They incur as two connected and important services residing inside the logical framework of the application. BaNaVA is designed using micro-services and React Native software framework, which allows the application to be developed cross-platform. This mobile application utilizes specific neural machine translation (NMT) and text-to-speech (TTS) technologies to efficiently operate within the edge computing environment of popular mobile devices.

   @inproceedings{thanh_banava_2023,
     title = {{BaNaVA}: {A} cross-platform {AI} mobile application for preserving the {Bahnaric} languages},
     copyright = {Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License},
     shorttitle = {{BaNaVA}},
     url = {https://ieeexplore.ieee.org/document/10466343},
     doi = {10.1109/BCD57833.2023.10466343},
     urldate = {2025-09-08},
     booktitle = {2023 {IEEE}/{ACIS} 8th {International} {Conference} on {Big} {Data}, {Cloud} {Computing}, and {Data} {Science} ({BCD})},
     author = {Quan, Tho T. and Dinh, Giang L. and Nguyen, Duc Q. and Vu, Hai H. and Nguyen, Quy T. and Tran, Khoa N. D. and Tran, Minh D.},
     month = oct,
     year = {2023},
     note = {ISSN: 2835-4419},
     keywords = {low-resource language, Data models, Data science, Barium, Linguistics, Machine translation, mobile application, Mobile applications, natural machine translation, Software, text-to-speech, voice conversion},
     pages = {120--125},
   }

2022

1. Sci. Rep.

   BeCaked: An Explainable Artificial Intelligence Model for COVID-19 Forecasting

   Duc Q. Nguyen, Nghia Q. Vo, Thinh T. Nguyen, and 4 more authors

   Scientific Reports, May 2022

   Publisher: Nature Publishing Group

   Abs DOI Bib PDF

   From the end of 2019, one of the most serious and largest spread pandemics occurred in Wuhan (China) named Coronavirus (COVID-19). As reported by the World Health Organization, there are currently more than 100 million infectious cases with an average mortality rate of about five percent all over the world. To avoid serious consequences on people’s lives and the economy, policies and actions need to be suitably made in time. To do that, the authorities need to know the future trend in the development process of this pandemic. This is the reason why forecasting models play an important role in controlling the pandemic situation. However, the behavior of this pandemic is extremely complicated and difficult to be analyzed, so that an effective model is not only considered on accurate forecasting results but also the explainable capability for human experts to take action pro-actively. With the recent advancement of Artificial Intelligence (AI) techniques, the emerging Deep Learning (DL) models have been proving highly effective when forecasting this pandemic future from the huge historical data. However, the main weakness of DL models is lacking the explanation capabilities. To overcome this limitation, we introduce a novel combination of the Susceptible-Infectious-Recovered-Deceased (SIRD) compartmental model and Variational Autoencoder (VAE) neural network known as BeCaked. With pandemic data provided by the Johns Hopkins University Center for Systems Science and Engineering, our model achieves 0.98 \\R^2\\and 0.012 MAPE at world level with 31-step forecast and up to 0.99 \\R^2\\and 0.0026 MAPE at country level with 15-step forecast on predicting daily infectious cases. Not only enjoying high accuracy, but BeCaked also offers useful justifications for its results based on the parameters of the SIRD model. Therefore, BeCaked can be used as a reference for authorities or medical experts to make on time right decisions.

   @article{nguyen_becaked_2022,
     title = {{BeCaked}: {An} {Explainable} {Artificial} {Intelligence} {Model} for {COVID}-19 {Forecasting}},
     volume = {12},
     copyright = {Creative Commons Attribution-ShareAlike 4.0 International License},
     issn = {2045-2322},
     shorttitle = {{BeCaked}},
     url = {https://www.nature.com/articles/s41598-022-11693-9},
     doi = {10.1038/s41598-022-11693-9},
     language = {en},
     number = {1},
     urldate = {2025-08-27},
     journal = {Scientific Reports},
     author = {Nguyen, Duc Q. and Vo, Nghia Q. and Nguyen, Thinh T. and Nguyen-An, Khuong and Nguyen, Quang H. and Tran, Dang N. and Quan, Tho T.},
     month = may,
     year = {2022},
     note = {Publisher: Nature Publishing Group},
     keywords = {Applied mathematics, Bioinformatics, Computational models, Computational science, Computer science, Differential equations, Disease model, Dynamical systems, Machine learning, Public health, Time series},
     pages = {7969},
   }

2. Social Bot Detector using Graph Neural Networks

   Hoang-Dung Nguyen, Duc Q. Nguyen, Hao Luong Pham, and 1 more author

   In 2022 RIVF International Conference on Computing and Communication Technologies (RIVF), Oct 2022

   ISSN: 2162-786X

   Abs DOI Bib PDF

   The importance of online social networks (OSNs) has been fueled by the human need for digital communication and broadcasting, as well as the improved state of internet connections and electronic devices. Meanwhile, social bots have been designed to automatically replicate the behavior of legitimate users in order to manipulate these OSNs. As a result, social bot detectors have been conducted concurrently, mostly on Twitter, in an attempt to discover new strategies for countering social bot attacks. In this paper, we propose SOBOG, a deep learning architecture that takes tweet relations, tweet semantics, and user properties into account to perform account-level and tweet-level detection. SOBOG also achieves outstanding performance on both tasks.

   @inproceedings{nguyen_social_2022,
     title = {Social {Bot} {Detector} using {Graph} {Neural} {Networks}},
     copyright = {Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License},
     url = {https://ieeexplore-ieee-org.libproxy1.nus.edu.sg/abstract/document/10013786},
     doi = {10.1109/RIVF55975.2022.10013786},
     urldate = {2025-09-08},
     booktitle = {2022 {RIVF} {International} {Conference} on {Computing} and {Communication} {Technologies} ({RIVF})},
     author = {Nguyen, Hoang-Dung and Nguyen, Duc Q. and Pham, Hao Luong and Tho, Quan Thanh},
     month = oct,
     year = {2022},
     note = {ISSN: 2162-786X},
     keywords = {Chatbots, Detectors, Feature extraction, graph neural networks, identity deception, information systems, Semantics, social bot, Social networking (online), Transformers, Vocabulary},
     pages = {197--202},
   }

3. Towards De Novo Drug Design for the Coronavirus: A Drug-Target Interaction Prediction Approach using Atom-enhanced Graph Neural Network with Multi-hop Gating Mechanism

   Duc Q. Nguyen, Khoan D. Le, Bach T. Ly, and 7 more authors

   In 2022 9th NAFOSTED Conference on Information and Computer Science (NICS), Oct 2022

   Best Paper Award Abs DOI Bib PDF

   Best Paper Award in NICS 2022

   For humans, the COVID-19 pandemic and Coronavirus have undeniably been a nightmare. Although there are effective vaccines, specific drugs are still urgent. Normally, to identify potential drugs, one needs to design and then test interactions between the drug and the virus in an in silico manner for determining candidates. This Drug-Target Interaction (DTI) process, can be done by molecular docking, which is too complicated and time-consuming for manual works. Therefore, it opens room for applying Artificial Intelligence (AI) techniques. In particular, Graph Neural Network (GNN) attracts recent attention since its high suitability for the nature of drug compounds and virus proteins. However, to introduce such a representation well-reflecting biological structures of biological compounds is not a trivial task. Moreover, since available datasets of Coronavirus are still not highly popular, the recently developed GNNs have been suffering from overfitting on them. We then address those issues by proposing a novel model known as Atom-enhanced Graph Neural Network with Multi-hop Gating Mechanism. On one hand, our model can learn more precise features of compounds and proteins. On the other hand, we introduce a new gating mechanism to create better atom representation from non-neighbor information. Once applying transfer learning from very large databanks, our model enjoys promising performance, especially when experimenting with Coronavirus.

   @inproceedings{nguyen_towards_2022,
     title = {Towards {De} {Novo} {Drug} {Design} for the {Coronavirus}: {A} {Drug}-{Target} {Interaction} {Prediction} {Approach} using {Atom}-enhanced {Graph} {Neural} {Network} with {Multi}-hop {Gating} {Mechanism}},
     copyright = {Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License},
     shorttitle = {Towards {De} {Novo} {Drug} {Design} for the {Coronavirus}},
     url = {https://ieeexplore.ieee.org/document/10013437},
     doi = {10.1109/NICS56915.2022.10013437},
     urldate = {2025-09-08},
     booktitle = {2022 9th {NAFOSTED} {Conference} on {Information} and {Computer} {Science} ({NICS})},
     author = {Nguyen, Duc Q. and Le, Khoan D. and Ly, Bach T. and Nguyen, An D. and Nguyen, Quang H. and Nguyen, Tuan H. and Quan, Tho T. and Duong, Cuong Quoc and Nguyen, Phuong Thuy Viet and Truong, Thanh N.},
     month = oct,
     year = {2022},
     keywords = {Coronavirus, COVID-19, Drug-Target Interaction, Drugs, Graph neural networks, Graph Neural Networks, Multi-hop Gating Mechanism, Proteins, Spread spectrum communication, Transfer learning, Vaccines},
     pages = {275--280},
   }

4. BeCaked+: An Explainable AI Model to Forecast Delta-Spreading Covid-19 Situations for Ho Chi Minh City

   Cuong Nguyen, Minh Nguyen, Duc Nguyen, and 4 more authors

   In Proceedings of the ICR’22 International Conference on Innovations in Computing Research, Oct 2022

   ISSN: 2194-5365

   Abs DOI Bib PDF

   Covid-19 is a global disaster that needs computing power to analyze, predict and interpret. So far, there have been several models doing the job. With a huge amount of daily data, deep learning models can be trained to achieve highly accurate forecasts but their...

   @inproceedings{nguyen_becaked_2022-1,
     title = {{BeCaked}+: {An} {Explainable} {AI} {Model} to {Forecast} {Delta}-{Spreading} {Covid}-19 {Situations} for {Ho} {Chi} {Minh} {City}},
     copyright = {All rights reserved},
     isbn = {978-3-031-14054-9},
     shorttitle = {{BeCaked}+},
     url = {https://link.springer.com/chapter/10.1007/978-3-031-14054-9_6},
     doi = {10.1007/978-3-031-14054-9_6},
     language = {en},
     urldate = {2025-09-08},
     booktitle = {Proceedings of the {ICR}’22 {International} {Conference} on {Innovations} in {Computing} {Research}},
     publisher = {Springer, Cham},
     author = {Nguyen, Cuong and Nguyen, Minh and Nguyen, Duc and Nguyen, Thinh and Nguyen-An, Khuong and Le, Chon and Quan, Tho},
     year = {2022},
     note = {ISSN: 2194-5365},
     pages = {53--64},
   }

2021

1. Realtime Bushfire Detection with Spatial-based Complex Event Processing

   Thanh Tam Nguyen, Thanh Toan Nguyen, Thanh Cong Phan, and 2 more authors

   In 2021 15th International Conference on Advanced Computing and Applications (ACOMP), Oct 2021

   ISSN: 2688-0202

   Abs DOI Bib PDF

   Bushfire is the primary destructive force that may cause damage to a large region, a country, or even the Earth. However, as bushfires spread too fast, they are often identified when they cannot control and cause significant damage. The reason is that existing works on remote sensing focus on low-level information processing, and thus, face the challenge of processing a massive amount of data in real-time. In this work, we employ complex event processing (CEP) to extract higher-level information to facilitate real-time bushfire detection. In particular, we propose a “spatial extension” to the ready-powerful CEP techniques to enable bushfire monitoring from the combinations of multiple spatial events. We further demonstrate the proposed spatial-based CEP on a real-time bushfire detection problem. Experimental results illustrate that our approach scales well while achieving the competitive detection performance.

   @inproceedings{nguyen_realtime_2021,
     title = {Realtime {Bushfire} {Detection} with {Spatial}-based {Complex} {Event} {Processing}},
     copyright = {Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License},
     url = {https://ieeexplore.ieee.org/document/9668230},
     doi = {10.1109/ACOMP53746.2021.00007},
     urldate = {2025-09-08},
     booktitle = {2021 15th {International} {Conference} on {Advanced} {Computing} and {Applications} ({ACOMP})},
     author = {Nguyen, Thanh Tam and Nguyen, Thanh Toan and Phan, Thanh Cong and Nguyen, Duc Q. and Nguyen, Quoc Viet Hung},
     month = oct,
     year = {2021},
     note = {ISSN: 2688-0202},
     keywords = {bushfire detection, complex event processing, Environmental monitoring, Force, Information processing, Measurement, Real-time systems, spatial query, Surface contamination, Water pollution},
     pages = {1--8},
   }

2020

1. ICCCI

   Automatic Container Code Recognition Using MultiDeep Pipeline

   Duy Nguyen, Duc Nguyen, Thong Nguyen, and 4 more authors

   In Advances in Computational Collective Intelligence, Oct 2020

   ISSN: 1865-0937

   Abs DOI Bib PDF

   Identification of license plates on intermodal containers (or containers) while entering and departing from the yard provides a wide range of practical benefits, such as organizing automatic opening of the rising arm barrier at the entrance and exit to and from the...

   @inproceedings{nguyen_automatic_2020,
     title = {Automatic {Container} {Code} {Recognition} {Using} {MultiDeep} {Pipeline}},
     copyright = {Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License},
     isbn = {978-3-030-63119-2},
     url = {https://link.springer.com/chapter/10.1007/978-3-030-63119-2_12},
     doi = {10.1007/978-3-030-63119-2_12},
     language = {en},
     urldate = {2025-09-08},
     booktitle = {Advances in {Computational} {Collective} {Intelligence}},
     publisher = {Springer, Cham},
     author = {Nguyen, Duy and Nguyen, Duc and Nguyen, Thong and Ngo, Khoi and Cao, Hung and Vuong, Thinh and Quan, Tho},
     year = {2020},
     note = {ISSN: 1865-0937},
     pages = {139--153},
   }

2. Design, implementation and evaluation for a high precision prosthetic hand using MyoBand and Random Forest algorithm

   Duc Q. Nguyen, Thien Cong Pham, and Tho Thanh Quan

   VNUHCM Journal of Engineering and Technology, Oct 2020

   Abs DOI Bib PDF

   A prosthesis is an equipment provided to people who lost one or some parts of their limbs to help them having almost normal behaviors in daily or hard activities. The convenience and intelligence of devices should create easiness and flexibility for users. Artificial devices require interdisciplinary collaboration from neurosurgeons, surgical surgeons, physiotherapists and equipment development. Computer engineering plays a crucial role in the design step, supporting manufacturing, training and recognition to match the desirability of customers. Moreover, users need a wide range of different options such as an aesthetic functional material, a myoelectric mechanism, a body-powered appliance or an activity specified device. Thus, the flexible configuration, the proper features and the cost are some important factors that drive user’s selection to the prosthesis. In this article, we describe an effective and powerful solution for analyzing, designing hardware and implementing software to train and recognize hand gestures for prosthetic arms. Moreover, we provide evaluation data of the method compared with similar approaches to support our design and implementation. This is fairly a complete system, making it a convenient solution for hand-cutoff people to control prosthetic hands using their electromyography signals. Statistical results with evaluations show that the device can respond correspondingly and the method creates promisingly recognition data after correct training processes. The prosthetic hardware implementation has also been simulated using a Light-emitting diode (LED) hand model with a high accuracy result.

   @article{nguyen_design_2020,
     title = {Design, implementation and evaluation for a high precision prosthetic hand using {MyoBand} and {Random} {Forest} algorithm},
     volume = {3},
     copyright = {Copyright (c) 2020 Duc Q. Nguyen, Thien Cong Pham, Tho Thanh Quan},
     issn = {2615-9872},
     url = {https://stdjet.scienceandtechnology.com.vn/index.php/stdjet/article/view/536},
     doi = {10.32508/stdjet.v3iSI1.536},
     language = {en},
     number = {SI1},
     urldate = {2025-09-08},
     journal = {VNUHCM Journal of Engineering and Technology},
     author = {Nguyen, Duc Q. and Pham, Thien Cong and Quan, Tho Thanh},
     month = oct,
     year = {2020},
     pages = {SI28--SI39},
   }