Yanglan Gan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Yanglan Gan has a total of 93 co-authors. Here are the top co-authors by number of shared publications:
- Guobing Zou
School of Computer Engineering and Science, Shanghai University, China
31 shared publications - Bofeng Zhang
Shanghai Polytechnic University
19 shared publications - Yixin Chen
Department of Computer Science and Engineering, Washington University in St. Louis, St. Louis, MO, USA
13 shared publications - Shengxiang Hu
School of Computer Engineering and Science, Shanghai University, Shanghai, China
10 shared publications - Song Yang
School of Computer Engineering and Science, Shanghai University, China
6 shared publications Guobing Zou
School of Computer Engineering and Science
5 shared publications
Latest Publications
- Open Access
- 17 November 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Knowledge Discovery From Data
Multi-level Collaborative Relation Graphs Enhanced LLM for Web Service Recommendation
Web Service (API) Recommendation enables intelligent selection of suitable APIs from large-scale repositories to compose mashup applications that satisfy complex user requirements. However, existing approaches face two critical limitations that affect further improvement in recommendation performance. First, the complementary synergy among multiple service relations has not been effectively exploited, particularly in cold-start scenarios where new mashups lack invocation history, making it difficult to obtain high-quality collaborative relation representations. Second, existing LLM-based approaches lack effective mechanisms for constructing and utilizing high-quality decision contexts specifically tailored for Web API recommendation, failing to fully unleash the potential of LLMs in Web API recommendation. To address these challenges, we propose LSFL, an LLM-enhanced Multi-level Service Collaborative Relation Graphs and Collaborative-Semantic Fusion Feature Learning framework. Specifically, LSFL constructs multi-level service collaborative relation graphs integrating invocation and category co-occurrence relations, and obtains high-quality collaborative relation embeddings through graph convolution and temperature-controlled attention. Furthermore, LSFL injects collaborative relation embeddings into the LLM as decision context via specially designed prompts, and employs Q-LoRA fine-tuning with a two-stage training strategy to extract deep collaborative-semantic fusion features for precise API recommendation. Extensive experiments against 15 baselines on a real-world dataset demonstrate that LSFL significantly outperforms existing state-of-the-art Web API recommendation approaches.
- 17 November 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Knowledge Discovery From Data
Open AccessMulti-level Collaborative Relation Graphs Enhanced LLM for Web Service Recommendation
Web Service (API) Recommendation enables intelligent selection of suitable APIs from large-scale repositories to compose mashup applications that satisfy complex user requirements. However, existing approaches face two critical limitations that affect further improvement in recommendation performance. First, the complementary synergy among multiple service relations has not been effectively exploited, particularly in cold-start scenarios where new mashups lack invocation history, making it difficult to obtain high-quality collaborative relation representations. Second, existing LLM-based approaches lack effective mechanisms for constructing and utilizing high-quality decision contexts specifically tailored for Web API recommendation, failing to fully unleash the potential of LLMs in Web API recommendation. To address these challenges, we propose LSFL, an LLM-enhanced Multi-level Service Collaborative Relation Graphs and Collaborative-Semantic Fusion Feature Learning framework. Specifically, LSFL constructs multi-level service collaborative relation graphs integrating invocation and category co-occurrence relations, and obtains high-quality collaborative relation embeddings through graph convolution and temperature-controlled attention. Furthermore, LSFL injects collaborative relation embeddings into the LLM as decision context via specially designed prompts, and employs Q-LoRA fine-tuning with a two-stage training strategy to extract deep collaborative-semantic fusion features for precise API recommendation. Extensive experiments against 15 baselines on a real-world dataset demonstrate that LSFL significantly outperforms existing state-of-the-art Web API recommendation approaches.
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13 (22) , 53204-53218
Spatio-Temporal Edge Data Caching Under Dynamic QoS Expectations and Fluctuating Server Loads
- 1 November 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13 (22) , 53204-53218
Spatio-Temporal Edge Data Caching Under Dynamic QoS Expectations and Fluctuating Server Loads
- Open Access
HierL2GDTA: Hierarchical Interaction Modeling with Local and Global Features for Drug–Target Affinity Prediction
Accurate prediction of drug–target binding affinity is essential for drug discovery. However, existing methods face limitations in capturing local and global molecular features and in effectively integrating protein and ligand representations for interaction modeling. Here, we present HierL2GDTA, a sequence-based hierarchical framework for predicting binding affinity. Pretrained protein and molecular representations are processed by multi-scale convolutions to extract local features, followed by a bidirectional gated recurrent unit and dilated convolutions to capture broader protein and ligand context, respectively. A modified Transformer encoder integrates global protein context and ligand features through sequential cross-attention to model intramolecular and intermolecular dependencies. HierL2GDTA outperformed the evaluated baseline methods on both BindingDB and Davis. Further evaluations demonstrated the effectiveness of the model components and favorable predictive performance across selected kinase families. Attention visualization highlighted molecular substructures associated with protein binding, while retrospective screening ranked known inhibitors of disease-associated targets among the leading candidates. These results indicate that HierL2GDTA is an effective deep learning method for binding affinity prediction, with potential application for target-based drug screening.
Open AccessHierL2GDTA: Hierarchical Interaction Modeling with Local and Global Features for Drug–Target Affinity Prediction
Accurate prediction of drug–target binding affinity is essential for drug discovery. However, existing methods face limitations in capturing local and global molecular features and in effectively integrating protein and ligand representations for interaction modeling. Here, we present HierL2GDTA, a sequence-based hierarchical framework for predicting binding affinity. Pretrained protein and molecular representations are processed by multi-scale convolutions to extract local features, followed by a bidirectional gated recurrent unit and dilated convolutions to capture broader protein and ligand context, respectively. A modified Transformer encoder integrates global protein context and ligand features through sequential cross-attention to model intramolecular and intermolecular dependencies. HierL2GDTA outperformed the evaluated baseline methods on both BindingDB and Davis. Further evaluations demonstrated the effectiveness of the model components and favorable predictive performance across selected kinase families. Attention visualization highlighted molecular substructures associated with protein binding, while retrospective screening ranked known inhibitors of disease-associated targets among the leading candidates. These results indicate that HierL2GDTA is an effective deep learning method for binding affinity prediction, with potential application for target-based drug screening.
- 8 June 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Services Computing
- Vol. PP (4) , 1-14
LMSR: LLM-enhanced Multi-perspective Service Feature Learning for Web API Recommendation
Web APIs have become a fundamental paradigm in the Web 4.0 era, with mashup services emerging as a transformative technology that combines multiple APIs to create comprehensive services. However, existing approaches exhibit two significant limitations: overlooking the quality and completeness of recommendation contexts of new mashup requirements, and failing to effectively extract high-quality collaborative features from multi-perspective service relationships. To address these limitations, we propose LMSR, a novel LLM-enhanced Multi-perspective Service Feature Learning framework for Web API Recommendation. LMSR first leverages general-purpose LLM to refine and encode the original requirement descriptions, and employs a Mixture of Service Experts (MoSE)-based context prediction model to precisely predict service information relevant to new requirements, establishing a comprehensive and high-quality recommendation context for mashup requirements. Furthermore, by integrating the predicted recommendation contexts into the LLM through fine-tuning, LMSR effectively extracts collaborative features from multi-perspective service relationships between mashup requirements and APIs, ultimately achieving precise Web API recommendation. Comprehensive experiments on real-world datasets demonstrate that LMSR significantly outperforms 11 baseline approaches across precision, recall, F1-score, and NDCG, validating its effectiveness in Web API recommendation. The codes are available at https://scdm-shu.github.io/codes/LMSR.zip.
- 8 June 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Services Computing
- Vol. PP (4) , 1-14
LMSR: LLM-enhanced Multi-perspective Service Feature Learning for Web API Recommendation
Web APIs have become a fundamental paradigm in the Web 4.0 era, with mashup services emerging as a transformative technology that combines multiple APIs to create comprehensive services. However, existing approaches exhibit two significant limitations: overlooking the quality and completeness of recommendation contexts of new mashup requirements, and failing to effectively extract high-quality collaborative features from multi-perspective service relationships. To address these limitations, we propose LMSR, a novel LLM-enhanced Multi-perspective Service Feature Learning framework for Web API Recommendation. LMSR first leverages general-purpose LLM to refine and encode the original requirement descriptions, and employs a Mixture of Service Experts (MoSE)-based context prediction model to precisely predict service information relevant to new requirements, establishing a comprehensive and high-quality recommendation context for mashup requirements. Furthermore, by integrating the predicted recommendation contexts into the LLM through fine-tuning, LMSR effectively extracts collaborative features from multi-perspective service relationships between mashup requirements and APIs, ultimately achieving precise Web API recommendation. Comprehensive experiments on real-world datasets demonstrate that LMSR significantly outperforms 11 baseline approaches across precision, recall, F1-score, and NDCG, validating its effectiveness in Web API recommendation. The codes are available at https://scdm-shu.github.io/codes/LMSR.zip.
- 26 February 2026
- Published by American Chemical Society (ACS) in Journal of Chemical Information and Modeling
- Vol. 66 (6) , 3320-3330
VFMol: A Discrete Flow Matching Variational Autoencoder for Molecular Graph Generation
Molecular graph generation is a key task in drug discovery, aiming to efficiently identify novel compounds with desired properties. While variational autoencoders (VAEs) excel at latent space modeling and discrete flow matching (DFM) enables efficient continuous-time sampling, existing approaches still face critical limitations. VAE decoders struggle with permutation invariance and suffer from one-shot generation bottlenecks, whereas DFM models often rely on a fixed prior initialization that lacks adaptability to specific molecular structures. To address these issues, we propose VFMol, a novel framework that synergistically integrates personalized VAE latent space modeling with the efficient stepwise sampling mechanism of DFM in the discrete space. Specifically, the encoder learns a posterior distribution tailored to each input graph as the generation starting point, thereby enhancing both structural fidelity and diversity. Moreover, we introduce a lightweight property-guided framework based on KAN and classifier-free guidance, enabling conditional generation without auxiliary property predictors. Experiments on two widely used molecular data sets demonstrate that VFMol achieves state-of-the-art performance in terms of molecular structural quality and property controllability, verifying its generality and effectiveness.
- 26 February 2026
- Published by American Chemical Society (ACS) in Journal of Chemical Information and Modeling
- Vol. 66 (6) , 3320-3330
VFMol: A Discrete Flow Matching Variational Autoencoder for Molecular Graph Generation
Molecular graph generation is a key task in drug discovery, aiming to efficiently identify novel compounds with desired properties. While variational autoencoders (VAEs) excel at latent space modeling and discrete flow matching (DFM) enables efficient continuous-time sampling, existing approaches still face critical limitations. VAE decoders struggle with permutation invariance and suffer from one-shot generation bottlenecks, whereas DFM models often rely on a fixed prior initialization that lacks adaptability to specific molecular structures. To address these issues, we propose VFMol, a novel framework that synergistically integrates personalized VAE latent space modeling with the efficient stepwise sampling mechanism of DFM in the discrete space. Specifically, the encoder learns a posterior distribution tailored to each input graph as the generation starting point, thereby enhancing both structural fidelity and diversity. Moreover, we introduce a lightweight property-guided framework based on KAN and classifier-free guidance, enabling conditional generation without auxiliary property predictors. Experiments on two widely used molecular data sets demonstrate that VFMol achieves state-of-the-art performance in terms of molecular structural quality and property controllability, verifying its generality and effectiveness.
- 22 February 2026
- Published by Elsevier in Knowledge-Based Systems
- Vol. 339
HINMOIT: Representation learning on heterogeneous information networks with multi-order information interaction
- 22 February 2026
- Published by Elsevier in Knowledge-Based Systems
- Vol. 339
HINMOIT: Representation learning on heterogeneous information networks with multi-order information interaction
- 18 February 2026
- Published by Elsevier in Knowledge-Based Systems
- Vol. 339
HASNN: Hierarchical attention spiking neural network for dynamic graph representation learning
- 18 February 2026
- Published by Elsevier in Knowledge-Based Systems
- Vol. 339
HASNN: Hierarchical attention spiking neural network for dynamic graph representation learning
- 30 December 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 307
Multi-view contrastive deep subspace clustering for attributed graph
- 30 December 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 307
Multi-view contrastive deep subspace clustering for attributed graph
- 1 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 30 (6) , 4932-4942
HCMAF: Hierarchical Feature Aggregation and Cross-Modal Attention Fusion Framework for Multi-Omics Patient Classification
The accumulation of large-scale multi-omics datasets has brought new opportunities for precise disease treatment. However, the inherent complexity of inter- and intra-omics relationships presents considerable obstacles to the precise integration of multi-omics data. Here, we propose a Hierarchical Feature Aggregation and Cross-Modal Attention Fusion (HCMAF) framework to integrate multi-omics data for patient classification and biomarker identification. Specifically, to capture both the specific information inherent in each omics data and complex cross-omics interactions, HCMAF incorporates three innovative modules. The hierarchical feature aggregation graph attention (HGAT) module captures intra-omics topological features through adaptive neighborhood aggregation. The cross-modal attention (CMA) module pinpoints inter-omics complementarity by modeling cross-omics dependencies. Finally, the confidence-driven multi-omics fusion (CMF) module dynamically integrates omics-specific predictions through learnable reliability weights. Comprehensive experiments on four public benchmark datasets show that HCMAF achieves better classification performance and consistently surpasses leading existing methods. Further component analysis confirms the crucial role of the HGAT, CMA and CMF modules in ensuring overall model effectiveness. HCMAF is implemented in Python using the Pytorch machine-learning library, and it is freely available at https://github.com/DHUDBlab/HCMAF.
- 1 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 30 (6) , 4932-4942
HCMAF: Hierarchical Feature Aggregation and Cross-Modal Attention Fusion Framework for Multi-Omics Patient Classification
The accumulation of large-scale multi-omics datasets has brought new opportunities for precise disease treatment. However, the inherent complexity of inter- and intra-omics relationships presents considerable obstacles to the precise integration of multi-omics data. Here, we propose a Hierarchical Feature Aggregation and Cross-Modal Attention Fusion (HCMAF) framework to integrate multi-omics data for patient classification and biomarker identification. Specifically, to capture both the specific information inherent in each omics data and complex cross-omics interactions, HCMAF incorporates three innovative modules. The hierarchical feature aggregation graph attention (HGAT) module captures intra-omics topological features through adaptive neighborhood aggregation. The cross-modal attention (CMA) module pinpoints inter-omics complementarity by modeling cross-omics dependencies. Finally, the confidence-driven multi-omics fusion (CMF) module dynamically integrates omics-specific predictions through learnable reliability weights. Comprehensive experiments on four public benchmark datasets show that HCMAF achieves better classification performance and consistently surpasses leading existing methods. Further component analysis confirms the crucial role of the HGAT, CMA and CMF modules in ensuring overall model effectiveness. HCMAF is implemented in Python using the Pytorch machine-learning library, and it is freely available at https://github.com/DHUDBlab/HCMAF.