Ha Duong Thi Hong
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Ha Duong Thi Hong has a total of 10 co-authors. Here are the top co-authors by number of shared publications:
- Vinh Truong Hoang
Ho Chi Minh City Open University, Ho Chi Minh City, Vietnam
4 shared publications - Bay Nguyen van
Faculty of Information Technology, Ho Chi Minh City Open University, Ho Chi Minh City, Vietnam
3 shared publications - Hau Nguyen Trung
Faculty of Information Technology, Ho Chi Minh City Open University, Ho Chi Minh City, Vietnam
2 shared publications Kiet Tran-Trung
Ho Chi Minh City Open University 97 Vo Van Tan , Xuan Hoa
2 shared publications- Kiet Tran-Trung
Ho Chi Minh City Open University
2 shared publications - Thien Ho Huong
Faculty of Information Technology, Ho Chi Minh City Open University, Ho Chi Minh City, Vietnam
2 shared publications
Latest Publications
- 13 August 2026
- Published by American Chemical Society (ACS) in Journal of Chemical Information and Modeling
- Vol. 66 (16) , 9787-9802
Dual-Attention Multimodal Framework for Molecular Property Prediction
Molecular property prediction plays a vital role in drug discovery and chemical research by facilitating the efficient selection and optimization of candidate compounds. Traditional Quantitative Structure–Activity Relationship models and early machine learning methods often rely on handcrafted molecular descriptors, which limit their generalizability and predictive performance. Recent advancements have shown that the powerful representation capabilities of Large Language Models on textual molecular data, combined with the relational learning strengths of Attention-based Graph Neural Networks, offer a new paradigm for capturing both the semantic and structural information about molecules. In this study, we introduce a novel Dual-Attention Multimodal framework for Graphs and Sequence-based representations, so-called DAM-GS. We evaluated our approach on multiple benchmark data sets covering diverse molecular prediction tasks. Experimental results showed that DAM-GS had outperformed the compared graph-based methods in most dataset–split settings. By leveraging the complementary advantages of graph-based and language-based modeling, our framework provides a promising solution for molecular property prediction with broad applications in drug discovery and computational molecular science.
- 13 August 2026
- Published by American Chemical Society (ACS) in Journal of Chemical Information and Modeling
- Vol. 66 (16) , 9787-9802
Dual-Attention Multimodal Framework for Molecular Property Prediction
Molecular property prediction plays a vital role in drug discovery and chemical research by facilitating the efficient selection and optimization of candidate compounds. Traditional Quantitative Structure–Activity Relationship models and early machine learning methods often rely on handcrafted molecular descriptors, which limit their generalizability and predictive performance. Recent advancements have shown that the powerful representation capabilities of Large Language Models on textual molecular data, combined with the relational learning strengths of Attention-based Graph Neural Networks, offer a new paradigm for capturing both the semantic and structural information about molecules. In this study, we introduce a novel Dual-Attention Multimodal framework for Graphs and Sequence-based representations, so-called DAM-GS. We evaluated our approach on multiple benchmark data sets covering diverse molecular prediction tasks. Experimental results showed that DAM-GS had outperformed the compared graph-based methods in most dataset–split settings. By leveraging the complementary advantages of graph-based and language-based modeling, our framework provides a promising solution for molecular property prediction with broad applications in drug discovery and computational molecular science.
- Open Access
- 5 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 159454-159486
Enhancing Smart Contract Security Through DevSecOps: An Adaptive Approach for Vulnerability Detection
Smart contracts are central to decentralized applications but are often vulnerable to security flaws, both during development and in deployment. Due to their immutable nature, any vulnerabilities introduced are permanent and exploitable, highlighting the need for a secure, structured development lifecycle. This study proposes an adaptive DevSecOps approach tailored to blockchain applications, drawing on established practices and gray literature. By embedding security into Continuous Integration and Continuous Delivery (CI/CD), the model ensures continuous protection from development through deployment. At its core is an adaptive cognitive security framework that automates vulnerability detection using techniques like static analysis, fuzzing, and symbolic execution. It also streamlines secure provisioning across various environments. To further enhance security, Agentic AI is integrated into the pipeline. These autonomous agents monitor code, infrastructure, and behavior in real-time, learning from past incidents, detecting threats, applying patches, and adapting CI/CD workflows dynamically. Experimental results validate the model’s effectiveness, showing significant improvements in the security and reliability of smart contracts. This approach offers a scalable, proactive solution for securing blockchain systems in a rapidly evolving threat landscape.
- 5 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 159454-159486
Open AccessEnhancing Smart Contract Security Through DevSecOps: An Adaptive Approach for Vulnerability Detection
Smart contracts are central to decentralized applications but are often vulnerable to security flaws, both during development and in deployment. Due to their immutable nature, any vulnerabilities introduced are permanent and exploitable, highlighting the need for a secure, structured development lifecycle. This study proposes an adaptive DevSecOps approach tailored to blockchain applications, drawing on established practices and gray literature. By embedding security into Continuous Integration and Continuous Delivery (CI/CD), the model ensures continuous protection from development through deployment. At its core is an adaptive cognitive security framework that automates vulnerability detection using techniques like static analysis, fuzzing, and symbolic execution. It also streamlines secure provisioning across various environments. To further enhance security, Agentic AI is integrated into the pipeline. These autonomous agents monitor code, infrastructure, and behavior in real-time, learning from past incidents, detecting threats, applying patches, and adapting CI/CD workflows dynamically. Experimental results validate the model’s effectiveness, showing significant improvements in the security and reliability of smart contracts. This approach offers a scalable, proactive solution for securing blockchain systems in a rapidly evolving threat landscape.
- Open Access
- 3 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 41422-41434
Water Meter Reading Based on Text Recognition Techniques and Deep Learning
In an increasingly technology-driven and automated world, optimizing energy management efficiency is crucial. A key component of this optimization is the precise measurement of water consumption, which plays a vital role in the allocation of resources for both households and businesses. While automated water meters are available, their adoption remains limited, particularly in less developed regions where manual meter reading is still common. This manual process often leads to inaccuracies and inefficiencies. This research aims to automate water meter readings, particularly in challenging and diverse environmental conditions. We propose a structured two-step approach to Automated Meter Reading (AMR), involving the identification of the counter region followed by digit segmentation and recognition. Using Convolutional Neural Networks (CNNs), particularly YOLOv8 for detecting the counter region, we evaluate six CNN-based techniques for digit segmentation and recognition: PP-OCRv3, Structure-Preserving Inner Offset Network (SPIN), Transformer-based Optical Character Recognition (TrOCR), RobustScanner, Show-Attend-and-Read (SAR), and Convolutional Recurrent Neural Network (CRNN).
- 3 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 41422-41434
Open AccessWater Meter Reading Based on Text Recognition Techniques and Deep Learning
In an increasingly technology-driven and automated world, optimizing energy management efficiency is crucial. A key component of this optimization is the precise measurement of water consumption, which plays a vital role in the allocation of resources for both households and businesses. While automated water meters are available, their adoption remains limited, particularly in less developed regions where manual meter reading is still common. This manual process often leads to inaccuracies and inefficiencies. This research aims to automate water meter readings, particularly in challenging and diverse environmental conditions. We propose a structured two-step approach to Automated Meter Reading (AMR), involving the identification of the counter region followed by digit segmentation and recognition. Using Convolutional Neural Networks (CNNs), particularly YOLOv8 for detecting the counter region, we evaluate six CNN-based techniques for digit segmentation and recognition: PP-OCRv3, Structure-Preserving Inner Offset Network (SPIN), Transformer-based Optical Character Recognition (TrOCR), RobustScanner, Show-Attend-and-Read (SAR), and Convolutional Recurrent Neural Network (CRNN).
- Open Access
- 26 October 2022
- Published by MDPI AG in Electronics
- Vol. 11 (21) , 3460
Weather Forecast Based on Color Cloud Image Recognition under the Combination of Local Image Descriptor and Histogram Selection
Numerous researchers have used machine vision in recent years to identify and categorize clouds according to their volume, shape, thickness, height, and coverage. Due to the significant variations in illumination, climate, and distortion that frequently characterize cloud images as a type of naturally striated structure, the Local Binary Patterns (LBP) descriptor and its variants have been proposed as feature extraction methods for characterizing natural texture images. Rotation invariance, low processing complexity, and resistance to monotonous brightness variations are characteristics of LBP. The disadvantage of LBP is that it produces binary data that are extremely noise-sensitive and it struggles on regions of the image that are “flat” because it depends on intensity differences. This paper considers the Local Ternary Patterns (LTP) feature to overcome the drawbacks of the LBP feature. We also propose the fusion of color characteristics, LBP features, and LTP features for the classification of cloud/sky images. Morover, this study proposes to apply the Intra-Class Similarity (ICS) technique, a histogram selection approach, with the goal of minimizing the number of histograms for characterizing images. The proposed approach achieves better performance of recognition with less features in use by fusing LBP and LTP features and using the ICS technique to choose potential histograms.
- 26 October 2022
- Published by MDPI AG in Electronics
- Vol. 11 (21) , 3460
Open AccessWeather Forecast Based on Color Cloud Image Recognition under the Combination of Local Image Descriptor and Histogram Selection
Numerous researchers have used machine vision in recent years to identify and categorize clouds according to their volume, shape, thickness, height, and coverage. Due to the significant variations in illumination, climate, and distortion that frequently characterize cloud images as a type of naturally striated structure, the Local Binary Patterns (LBP) descriptor and its variants have been proposed as feature extraction methods for characterizing natural texture images. Rotation invariance, low processing complexity, and resistance to monotonous brightness variations are characteristics of LBP. The disadvantage of LBP is that it produces binary data that are extremely noise-sensitive and it struggles on regions of the image that are “flat” because it depends on intensity differences. This paper considers the Local Ternary Patterns (LTP) feature to overcome the drawbacks of the LBP feature. We also propose the fusion of color characteristics, LBP features, and LTP features for the classification of cloud/sky images. Morover, this study proposes to apply the Intra-Class Similarity (ICS) technique, a histogram selection approach, with the goal of minimizing the number of histograms for characterizing images. The proposed approach achieves better performance of recognition with less features in use by fusing LBP and LTP features and using the ICS technique to choose potential histograms.