Hongming Zhu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Hongming Zhu has a total of 76 co-authors. Here are the top co-authors by number of shared publications:
- Qin Liu
School of Computer Science and Technology, Tongji University, Shanghai, China
15 shared publications - Bowen Du
Tongji University, Shanghai, China
9 shared publications - Xiaoguo Wang
School of Computer Science and Technology, Tongji University, Shanghai, China
7 shared publications - Sicong Liu
Tongji University
6 shared publications - Xiaowen Wang
Tongji University
6 shared publications - Chao Chen
Data Management and Application Department, Bank of Shanghai, Shanghai, China
4 shared publications
Latest Publications
- 19 July 2026
- Published by Springer Nature
- p. 266-278
A Two-Stage Balanced Learning Framework for Unbiased Scene Graph Generation
- 19 July 2026
- Published by Springer Nature
- p. 266-278
A Two-Stage Balanced Learning Framework for Unbiased Scene Graph Generation
- Open Access
- 15 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 397-403
An Intent Detection and Slot Filling Model for Financial Multi-Turn Dialogues
- 15 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 397-403
Open AccessAn Intent Detection and Slot Filling Model for Financial Multi-Turn Dialogues
- 29 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 64 (01962892) , 4408515
CAB-SCD: Semantic Change Detection Based on Cost Aggregation
Semantic change detection (SCD), which involves simultaneously localizing changed areas and categorizing their semantic shifts, plays a critical role in monitoring the Earth’s surface. Most existing SCD methods adopt a multitask learning framework that decomposes SCD into a binary change detection (BCD) sub-task and dual semantic segmentation (SS) sub-tasks. However, these approaches often overemphasize visual representations while ignoring the rich semantic context provided by textual data, hindering their robustness against intra-class variation and inter-class similarity in remote sensing imagery. They also encounter challenges, such as loss of spatial details caused by deficient multiscale feature fusion, and limited change perception due to shared feature learning, resulting in fragmented segmentation and rough boundaries. Inspired by recent advances in vision language models (VLMs), we propose a novel framework called CAB-SCD, to alleviate these issues. Firstly, we design a CLIP-based cost map generation module (CMG), integrating visual data with textual descriptions to enrich feature encoding. Secondly, the spatial and class refinement-based upsampling module (SCRU) is developed to produce organized segmentation regions through progressive multiscale fusion. Finally, the difference-aware fusion module (DAF) aims to learn dedicated change-specific features and helps improve sensitivity to diverse changes. Experimental results show that our model surpasses existing state-of-the-art methods on three SCD datasets, demonstrating that integrating VLMs’ textual priors and explicit change modeling into the process of multiscale fusion can lead to superior performance. Our code is available at https://github.com/Luuccia/CAB-SCD.
- 29 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 64 (01962892) , 4408515
CAB-SCD: Semantic Change Detection Based on Cost Aggregation
Semantic change detection (SCD), which involves simultaneously localizing changed areas and categorizing their semantic shifts, plays a critical role in monitoring the Earth’s surface. Most existing SCD methods adopt a multitask learning framework that decomposes SCD into a binary change detection (BCD) sub-task and dual semantic segmentation (SS) sub-tasks. However, these approaches often overemphasize visual representations while ignoring the rich semantic context provided by textual data, hindering their robustness against intra-class variation and inter-class similarity in remote sensing imagery. They also encounter challenges, such as loss of spatial details caused by deficient multiscale feature fusion, and limited change perception due to shared feature learning, resulting in fragmented segmentation and rough boundaries. Inspired by recent advances in vision language models (VLMs), we propose a novel framework called CAB-SCD, to alleviate these issues. Firstly, we design a CLIP-based cost map generation module (CMG), integrating visual data with textual descriptions to enrich feature encoding. Secondly, the spatial and class refinement-based upsampling module (SCRU) is developed to produce organized segmentation regions through progressive multiscale fusion. Finally, the difference-aware fusion module (DAF) aims to learn dedicated change-specific features and helps improve sensitivity to diverse changes. Experimental results show that our model surpasses existing state-of-the-art methods on three SCD datasets, demonstrating that integrating VLMs’ textual priors and explicit change modeling into the process of multiscale fusion can lead to superior performance. Our code is available at https://github.com/Luuccia/CAB-SCD.
- 27 April 2026
- Published by Elsevier in Computers & Geosciences
- Vol. 213
DGSeg: Dual guidance with textual priors and structural awareness for open-vocabulary remote sensing segmentation
- 27 April 2026
- Published by Elsevier in Computers & Geosciences
- Vol. 213
DGSeg: Dual guidance with textual priors and structural awareness for open-vocabulary remote sensing segmentation
- Open Access
- 6 January 2026
- Published by Springer Nature in BMC Bioinformatics
- Vol. 27 (1)
Hgtsynergy: a transfer learning method for predicting anticancer synergistic drug combinations based on a drug-drug interaction heterogeneous graph
Background: Drug combination therapy often outperforms monotherapy in cancer treatment, but the vast number of available drugs makes manual screening for synergistic combinations costly. Computational methods, especially deep learning, can reduce the search space by predicting likely synergistic drug combinations. Recent studies have improved drug synergy prediction by modeling associations among different biological entities, but drug–drug interactions have not been fully leveraged in this scenario, which motivated the work presented in this paper. Methods: This paper proposes a deep learning method named HGTSynergy to predict synergistic drug combinations, which employs a heterogeneous graph attention network and a tailored task to capture complex latent patterns in the drug network as prior knowledge. The learned knowledge is then transferred through a transfer learning framework to the downstream task of predicting drug synergy scores, effectively enhancing predictive performance. Results: A five-fold nested cross-validation is employed to train HGTSynergy. In the synergy regression task, HGTSynergy outperforms seven deep learning methods, achieving a mean squared error of 222.83, root mean squared error of 14.91, and Pearson correlation coefficient of 0.75. For the synergy classification task, it also surpasses other methods with an area under the receiver operating characteristic curve of 0.90, area under the precision–recall curve of 0.63, accuracy of 0.94, precision of 0.72, and Cohen’s Kappa of 0.52. The ablation study verifies that the heterogeneous graph attention network and the transfer learning framework both have a positive effect on prediction performance. Moreover, a series of analyses demonstrates that the proposed method exhibits strong generalization performance and interpretability. The case study further validates its consistency with prior research. Conclusions: This study suggests that drug synergy prediction can be improved by comprehensively modeling diverse drug–drug interaction types and leveraging transfer learning to extract prior knowledge from them. The ability of HGTSynergy to discover new anticancer synergistic drug combinations outperforms other state-of-the-art methods. HGTSynergy promises to be a powerful tool to pre-screen anticancer synergistic drug combinations.
- 6 January 2026
- Published by Springer Nature in BMC Bioinformatics
- Vol. 27 (1)
Open AccessHgtsynergy: a transfer learning method for predicting anticancer synergistic drug combinations based on a drug-drug interaction heterogeneous graph
Background: Drug combination therapy often outperforms monotherapy in cancer treatment, but the vast number of available drugs makes manual screening for synergistic combinations costly. Computational methods, especially deep learning, can reduce the search space by predicting likely synergistic drug combinations. Recent studies have improved drug synergy prediction by modeling associations among different biological entities, but drug–drug interactions have not been fully leveraged in this scenario, which motivated the work presented in this paper. Methods: This paper proposes a deep learning method named HGTSynergy to predict synergistic drug combinations, which employs a heterogeneous graph attention network and a tailored task to capture complex latent patterns in the drug network as prior knowledge. The learned knowledge is then transferred through a transfer learning framework to the downstream task of predicting drug synergy scores, effectively enhancing predictive performance. Results: A five-fold nested cross-validation is employed to train HGTSynergy. In the synergy regression task, HGTSynergy outperforms seven deep learning methods, achieving a mean squared error of 222.83, root mean squared error of 14.91, and Pearson correlation coefficient of 0.75. For the synergy classification task, it also surpasses other methods with an area under the receiver operating characteristic curve of 0.90, area under the precision–recall curve of 0.63, accuracy of 0.94, precision of 0.72, and Cohen’s Kappa of 0.52. The ablation study verifies that the heterogeneous graph attention network and the transfer learning framework both have a positive effect on prediction performance. Moreover, a series of analyses demonstrates that the proposed method exhibits strong generalization performance and interpretability. The case study further validates its consistency with prior research. Conclusions: This study suggests that drug synergy prediction can be improved by comprehensively modeling diverse drug–drug interaction types and leveraging transfer learning to extract prior knowledge from them. The ability of HGTSynergy to discover new anticancer synergistic drug combinations outperforms other state-of-the-art methods. HGTSynergy promises to be a powerful tool to pre-screen anticancer synergistic drug combinations.
- Open Access
- 14 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 57-64
Graph-Augmented Contrastive Networks for Predicting Customer Product Preferences in Bank Services
- 14 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 57-64
Open AccessGraph-Augmented Contrastive Networks for Predicting Customer Product Preferences in Bank Services
- Open Access
- 14 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 65-73
Temporal Multi-scale Contrastive Learning for Credit Card Customer Dormancy Prediction
- 14 November 2025
- Published by Association for Computing Machinery (ACM)
- p. 65-73
Open AccessTemporal Multi-scale Contrastive Learning for Credit Card Customer Dormancy Prediction
- 8 September 2025
- Published by World Scientific Pub Co Pte Ltd in International Journal of Software Engineering and Knowledge Engineering
- Vol. 35 (11) , 1669-1684
STGAN-CR: A Semantics-Aware Cloud Removal Network Integrating Swin Transformer and GANs for Remote Sensing Applications
Advancements in satellite remote sensing have enhanced Earth observation, enabling the acquisition of images crucial for various applications. However, cloud cover degrades image quality and obstructs both low-level details and high-level semantic information, impacting downstream tasks like scene classification. Existing methods often fail to balance these features, particularly under complex conditions. To address these challenges, we propose STGAN-CR, a semantics-aware cloud removal framework that integrates Swin Transformer [ 1 ] with Generative Adversarial Networks (GANs) to optimize both detail recovery and semantic preservation. Leveraging the Swin Transformer’s global modeling capabilities, STGAN-CR excels in capturing complex spatial dependencies and restoring semantically meaningful features. To better evaluate practical utility, we introduce a novel evaluation metric based on scene classification accuracy, directly reflecting the restoration of high-level semantic information. Extensive experiments and ablation studies demonstrate that STGAN-CR outperforms existing models in both visual quality and semantic understanding, providing robust de-clouded images for downstream remote sensing applications.
- 8 September 2025
- Published by World Scientific Pub Co Pte Ltd in International Journal of Software Engineering and Knowledge Engineering
- Vol. 35 (11) , 1669-1684
STGAN-CR: A Semantics-Aware Cloud Removal Network Integrating Swin Transformer and GANs for Remote Sensing Applications
Advancements in satellite remote sensing have enhanced Earth observation, enabling the acquisition of images crucial for various applications. However, cloud cover degrades image quality and obstructs both low-level details and high-level semantic information, impacting downstream tasks like scene classification. Existing methods often fail to balance these features, particularly under complex conditions. To address these challenges, we propose STGAN-CR, a semantics-aware cloud removal framework that integrates Swin Transformer [ 1 ] with Generative Adversarial Networks (GANs) to optimize both detail recovery and semantic preservation. Leveraging the Swin Transformer’s global modeling capabilities, STGAN-CR excels in capturing complex spatial dependencies and restoring semantically meaningful features. To better evaluate practical utility, we introduce a novel evaluation metric based on scene classification accuracy, directly reflecting the restoration of high-level semantic information. Extensive experiments and ablation studies demonstrate that STGAN-CR outperforms existing models in both visual quality and semantic understanding, providing robust de-clouded images for downstream remote sensing applications.
- Open Access
- 25 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 457-463
Memory-Enhanced Multi-Head Self-Attention Feature Ranking Model for Financial Data Fluctuations
- 25 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 457-463
Open AccessMemory-Enhanced Multi-Head Self-Attention Feature Ranking Model for Financial Data Fluctuations
- Open Access
- 25 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 464-470
A Multi-Factor-Based Feature Ranking Model for Financial Data Fluctuations
- 25 July 2025
- Published by Association for Computing Machinery (ACM)
- p. 464-470
Open AccessA Multi-Factor-Based Feature Ranking Model for Financial Data Fluctuations