Junyu Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Junyu Li has a total of 37 co-authors. Here are the top co-authors by number of shared publications:
- Hongmin Cai
South China University of Technology, Guangzhou, China
12 shared publications - Fei Qi
Guizhou Minzu University
7 shared publications - Jiazhou Chen
School of Computer Science and Technology, Guangdong University of Technology, Guangzhou, China
5 shared publications - Bin Hu
Key Laboratory of Brain Health Intelligent Evaluation and Intervention, Ministry of Education, Beijing Institute of Technology, Beijing, China
3 shared publications - Bin Zhang
Guangdong Institute of Intelligence Science and Technology, Hengqin, China
3 shared publications - Weitian Huang
School of Computer Science and Engineering, South China University of Technology, Guangzhou, China
3 shared publications
Latest Publications
- Open Access
- 17 September 2026
- Published by Elsevier in Artificial Intelligence
- Vol. 361
SiameseNet+: Classification of HDLSS data with tensor affinities
- 17 September 2026
- Published by Elsevier in Artificial Intelligence
- Vol. 361
Open AccessSiameseNet+: Classification of HDLSS data with tensor affinities
- Open Access
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 10613-10622
ENGGT: An Edge-Node Guided Graph Transformer for Spatial Domain Parsing with Accurate Boundaries
- 8 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 10613-10622
Open AccessENGGT: An Edge-Node Guided Graph Transformer for Spatial Domain Parsing with Accurate Boundaries
- 12 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 30 (7) , 5783-5793
GraphSTAR: Proximal Operator-Based Graph Neural Network Enhanced by Dynamic Graph Aggregation for Spatial Transcriptomics
Spatial transcriptomics technologies carry out advanced sequencing analysis of molecular profiles with a spatial context, providing multi-source information essential for elucidating biological regulatory mechanisms. Nonetheless, it poses challenges in the integration of raw spatial coordinates with high-dimensional gene expression profiles in their native feature space. While spatial-aware methods effectively aggregate molecular information from local spatial neighborhoods, they fail to explore the long-range relationships associated with gene expression data. To address this issue, this paper introduces a novel approach termed GraphSTAR that encodes both spatial and gene expression data into undirected graphs, characterizing the local spatial proximity and global transcriptional similarity, respectively. Through a graph aggregation process, GraphSTAR integrates these diverse data sources within a joint graph structure, effectively modeling both local neighborhood relationships and long-range functional associations. Subsequently, a reassembled graph neural network is established by incorporating the graph aggregation into the feed-forward propagation using proximal operators, progressively refining spatial-informed latent representation to decipher spatial expression patterns of genes. Extensive experiments on benchmark datasets demonstrate that GraphSTAR outperforms state-of-the-art methods in both spatial domain identification and cell-type annotation tasks.
- 12 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 30 (7) , 5783-5793
GraphSTAR: Proximal Operator-Based Graph Neural Network Enhanced by Dynamic Graph Aggregation for Spatial Transcriptomics
Spatial transcriptomics technologies carry out advanced sequencing analysis of molecular profiles with a spatial context, providing multi-source information essential for elucidating biological regulatory mechanisms. Nonetheless, it poses challenges in the integration of raw spatial coordinates with high-dimensional gene expression profiles in their native feature space. While spatial-aware methods effectively aggregate molecular information from local spatial neighborhoods, they fail to explore the long-range relationships associated with gene expression data. To address this issue, this paper introduces a novel approach termed GraphSTAR that encodes both spatial and gene expression data into undirected graphs, characterizing the local spatial proximity and global transcriptional similarity, respectively. Through a graph aggregation process, GraphSTAR integrates these diverse data sources within a joint graph structure, effectively modeling both local neighborhood relationships and long-range functional associations. Subsequently, a reassembled graph neural network is established by incorporating the graph aggregation into the feed-forward propagation using proximal operators, progressively refining spatial-informed latent representation to decipher spatial expression patterns of genes. Extensive experiments on benchmark datasets demonstrate that GraphSTAR outperforms state-of-the-art methods in both spatial domain identification and cell-type annotation tasks.
- Open Access
- 26 December 2025
- Published by Oxford University Press (OUP) in Genomics, Proteomics & Bioinformatics
- Vol. 24 (1)
HyperSTAR: Unveiling Tissue Structure and Tumor Microenvironment from Spatial Omics by Hypergraph Learning
Spatial omics technologies have revolutionized life sciences by enabling the simultaneous acquisition of biomolecular and spatial information. Identifying spatial patterns is crucial for understanding organ development and tumor microenvironments. However, the emergence of diverse spatial omics resolutions in these technologies has made it challenging to accurately characterize spatial domains at finer resolutions. To address this, we propose HyperSTAR, a hypergraph-based method designed to precisely identify spatial domains across varying resolutions by leveraging higher-order relationships among spatially adjacent tissue programs. Specifically, a gene expression-guided hyperedge decomposition module is introduced to refine the hypergraph structure to accurately delineate spatial domain boundaries. Additionally, a hypergraph attention convolutional neural network is designed to adaptively learn the importance of each hyperedge, enhancing the model’s ability to capture complex higher-order relationships within spatially neighboring multi-spots and/or single cells. HyperSTAR outperforms existing graph neural network models in tasks such as uncovering tissue substructures, inferring spatiotemporal patterns, and denoising spatially resolved gene expression. It effectively handles diverse spatial omics data types and scales seamlessly to large datasets. The method successfully reveals spatial heterogeneity in breast cancer sections, with findings validated through functional and survival analyses of independent clinical data. HyperSTAR represents a significant advancement in spatial omics analysis, representing a robust tool for exploring complex spatial patterns across varying resolutions and data types. Its ability to capture intricate higher-order relationships among spatially neighboring spots/cells makes it an invaluable tool for advancing research in life sciences, particularly in cancer and developmental biology. The toolbox is available at https://github.com/Ringoio/HyperSTAR.
- 26 December 2025
- Published by Oxford University Press (OUP) in Genomics, Proteomics & Bioinformatics
- Vol. 24 (1)
Open AccessHyperSTAR: Unveiling Tissue Structure and Tumor Microenvironment from Spatial Omics by Hypergraph Learning
Spatial omics technologies have revolutionized life sciences by enabling the simultaneous acquisition of biomolecular and spatial information. Identifying spatial patterns is crucial for understanding organ development and tumor microenvironments. However, the emergence of diverse spatial omics resolutions in these technologies has made it challenging to accurately characterize spatial domains at finer resolutions. To address this, we propose HyperSTAR, a hypergraph-based method designed to precisely identify spatial domains across varying resolutions by leveraging higher-order relationships among spatially adjacent tissue programs. Specifically, a gene expression-guided hyperedge decomposition module is introduced to refine the hypergraph structure to accurately delineate spatial domain boundaries. Additionally, a hypergraph attention convolutional neural network is designed to adaptively learn the importance of each hyperedge, enhancing the model’s ability to capture complex higher-order relationships within spatially neighboring multi-spots and/or single cells. HyperSTAR outperforms existing graph neural network models in tasks such as uncovering tissue substructures, inferring spatiotemporal patterns, and denoising spatially resolved gene expression. It effectively handles diverse spatial omics data types and scales seamlessly to large datasets. The method successfully reveals spatial heterogeneity in breast cancer sections, with findings validated through functional and survival analyses of independent clinical data. HyperSTAR represents a significant advancement in spatial omics analysis, representing a robust tool for exploring complex spatial patterns across varying resolutions and data types. Its ability to capture intricate higher-order relationships among spatially neighboring spots/cells makes it an invaluable tool for advancing research in life sciences, particularly in cancer and developmental biology. The toolbox is available at https://github.com/Ringoio/HyperSTAR.
- 24 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 47 (1) , 387-399
Unsupervised Dual Deep Hashing With Semantic-Index and Content-Code for Cross-Modal Retrieval
Hashing technology has exhibited great cross-modal retrieval potential due to its appealing retrieval efficiency and storage effectiveness. Most current supervised cross-modal retrieval methods heavily rely on accurate semantic supervision, which is intractable for annotations with ever-growing sample sizes. By comparison, the existing unsupervised methods rely on accurate sample similarity preservation strategies with intensive computational costs to compensate for the lack of semantic guidance, which causes these methods to lose the power to bridge the semantic gap. Furthermore, both kinds of approaches need to search for the nearest samples among all samples in a large search space, whose process is laborious. To address these issues, this paper proposes an unsupervised dual deep hashing (UDDH) method with semantic-index and content-code for cross-modal retrieval. Deep hashing networks are utilized to extract deep features and jointly encode the dual hashing codes in a collaborative manner with a common semantic index and modality content codes to simultaneously bridge the semantic and heterogeneous gaps for cross-modal retrieval. The dual deep hashing architecture, comprising the head code on semantic index and tail codes on modality content, enhances the efficiency for cross-modal retrieval. A query sample only needs to search for the retrieved samples with the same semantic index, thus greatly shrinking the search space and achieving superior retrieval efficiency. UDDH integrates the learning processes of deep feature extraction, binary optimization, common semantic index, and modality content code within a unified model, allowing for collaborative optimization to enhance the overall performance. Extensive experiments are conducted to demonstrate the retrieval superiority of the proposed approach over the state-of-the-art baselines
- 24 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 47 (1) , 387-399
Unsupervised Dual Deep Hashing With Semantic-Index and Content-Code for Cross-Modal Retrieval
Hashing technology has exhibited great cross-modal retrieval potential due to its appealing retrieval efficiency and storage effectiveness. Most current supervised cross-modal retrieval methods heavily rely on accurate semantic supervision, which is intractable for annotations with ever-growing sample sizes. By comparison, the existing unsupervised methods rely on accurate sample similarity preservation strategies with intensive computational costs to compensate for the lack of semantic guidance, which causes these methods to lose the power to bridge the semantic gap. Furthermore, both kinds of approaches need to search for the nearest samples among all samples in a large search space, whose process is laborious. To address these issues, this paper proposes an unsupervised dual deep hashing (UDDH) method with semantic-index and content-code for cross-modal retrieval. Deep hashing networks are utilized to extract deep features and jointly encode the dual hashing codes in a collaborative manner with a common semantic index and modality content codes to simultaneously bridge the semantic and heterogeneous gaps for cross-modal retrieval. The dual deep hashing architecture, comprising the head code on semantic index and tail codes on modality content, enhances the efficiency for cross-modal retrieval. A query sample only needs to search for the retrieved samples with the same semantic index, thus greatly shrinking the search space and achieving superior retrieval efficiency. UDDH integrates the learning processes of deep feature extraction, binary optimization, common semantic index, and modality content code within a unified model, allowing for collaborative optimization to enhance the overall performance. Extensive experiments are conducted to demonstrate the retrieval superiority of the proposed approach over the state-of-the-art baselines
- 19 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Emerging Topics in Computational Intelligence
- Vol. 9 (1) , 406-418
Tensorlized Multi-Kernel Clustering via Consensus Tensor Decomposition
Multi-kernel clustering aims to learn a fused kernel from a set of base kernels. However, conventional multi-kernel clustering methods typically suffer from inherent limitations in exploiting the interrelations and complementarity between the kernels. The noises and redundant information from original base kernels also lead to contamination of the fused kernel. To address these issues, this paper presents a Tensorlized Multi-Kernel Clustering (TensorMKC) method. The proposed TensorMKC stacks kernel matrices into a kernel tensor along the kernel space. To attain consensus extraction while mitigating the impact of noise, we incorporate the tensor low-rank constraint into the process of learning base kernels. Subsequently, a tensor-based weighted fusion strategy is employed to integrate the refined base kernels, yielding an optimized fused kernel for clustering. The process of kernel learning is formulated as a joint minimization problem to seek the promising fusion solution. Through extensive comparative experiments with fifteen popular methods on ten benchmark datasets from various fields, the results demonstrate that TensorMKC exhibits superior performance.
- 19 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Emerging Topics in Computational Intelligence
- Vol. 9 (1) , 406-418
Tensorlized Multi-Kernel Clustering via Consensus Tensor Decomposition
Multi-kernel clustering aims to learn a fused kernel from a set of base kernels. However, conventional multi-kernel clustering methods typically suffer from inherent limitations in exploiting the interrelations and complementarity between the kernels. The noises and redundant information from original base kernels also lead to contamination of the fused kernel. To address these issues, this paper presents a Tensorlized Multi-Kernel Clustering (TensorMKC) method. The proposed TensorMKC stacks kernel matrices into a kernel tensor along the kernel space. To attain consensus extraction while mitigating the impact of noise, we incorporate the tensor low-rank constraint into the process of learning base kernels. Subsequently, a tensor-based weighted fusion strategy is employed to integrate the refined base kernels, yielding an optimized fused kernel for clustering. The process of kernel learning is formulated as a joint minimization problem to seek the promising fusion solution. Through extensive comparative experiments with fifteen popular methods on ten benchmark datasets from various fields, the results demonstrate that TensorMKC exhibits superior performance.
- Open Access
- 14 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks and Learning Systems
- Vol. 36 (5) , 8699-8713
Uniform Tensor Clustering by Jointly Exploring Sample Affinities of Various Orders
Traditional clustering methods rely on pairwise affinity to divide samples into different subgroups. However, high-dimensional small-sample (HDLSS) data are affected by the concentration effects, rendering traditional pairwise metrics unable to accurately describe relationships between samples, leading to suboptimal clustering results. This article advances the proposition of employing high-order affinities to characterize multiple sample relationships as a strategic means to circumnavigate the concentration effects. We establish a nexus between different order affinities by constructing specialized decomposable high-order affinities, thereby formulating a uniform mathematical framework. Building upon this insight, a novel clustering method named uniform tensor clustering (UTC) is proposed, which learns a consensus low-dimensional embedding for clustering by the synergistic exploitation of multiple-order affinities. Extensive experiments on synthetic and real-world datasets demonstrate two findings: 1) high-order affinities are better suited for characterizing sample relationships in complex data and 2) reasonable use of different order affinities can enhance clustering effectiveness, especially in handling high-dimensional data.
- 14 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks and Learning Systems
- Vol. 36 (5) , 8699-8713
Open AccessUniform Tensor Clustering by Jointly Exploring Sample Affinities of Various Orders
Traditional clustering methods rely on pairwise affinity to divide samples into different subgroups. However, high-dimensional small-sample (HDLSS) data are affected by the concentration effects, rendering traditional pairwise metrics unable to accurately describe relationships between samples, leading to suboptimal clustering results. This article advances the proposition of employing high-order affinities to characterize multiple sample relationships as a strategic means to circumnavigate the concentration effects. We establish a nexus between different order affinities by constructing specialized decomposable high-order affinities, thereby formulating a uniform mathematical framework. Building upon this insight, a novel clustering method named uniform tensor clustering (UTC) is proposed, which learns a consensus low-dimensional embedding for clustering by the synergistic exploitation of multiple-order affinities. Extensive experiments on synthetic and real-world datasets demonstrate two findings: 1) high-order affinities are better suited for characterizing sample relationships in complex data and 2) reasonable use of different order affinities can enhance clustering effectiveness, especially in handling high-dimensional data.
- 8 April 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 26 (15209210) , 9109-9120
Unsupervised Dual Hashing Coding (UDC) on Semantic Tagging and Sample Content for Cross-Modal Retrieval
Current cross-modal retrieval methods heavily rely on accurate semantic labels or sample similarity measurements, and need to search for the nearest samples among all samples in the huge search space, severely limiting the application in stratifying large-scale and high-dimensional multimodal data. To tackle with the issues, this paper proposes an unsupervised cross-modal retrieval method to bypass the semanticwise supervision and samplewise similarity from a standpoint of featurewise matching, named by unsupervised dual hashing coding (UDC). It jointly learns the dual hashing codes on semantic tagging and sample content through factorizing a feature matching potential, which is allowed to bridge the semantic and heterogeneous gaps among different modalities simultaneously through maintaining the inter-modality-consistent semantic information and cross-modality-correlated sample content. In this way, each sample is uniquely coded by a head code on semanticwise tags, and tail codes on samplewise content. The dual coding design makes it very efficient for sample retrieval, in which the query sample only need to search for the retrieved ones with the same semantic tag, greatly narrowing down the search space. The proposed model avoids the calculation of massive sample-wise similarity and works with dual hashing coding scheme, which achieves a twofold efficiency enhancement for analyzing the large-scale and high-dimensional multimodal data. Extensive experiments have been conducted to demonstrate that it achieved superiority on computational time and retrieval performance.
- 8 April 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 26 (15209210) , 9109-9120
Unsupervised Dual Hashing Coding (UDC) on Semantic Tagging and Sample Content for Cross-Modal Retrieval
Current cross-modal retrieval methods heavily rely on accurate semantic labels or sample similarity measurements, and need to search for the nearest samples among all samples in the huge search space, severely limiting the application in stratifying large-scale and high-dimensional multimodal data. To tackle with the issues, this paper proposes an unsupervised cross-modal retrieval method to bypass the semanticwise supervision and samplewise similarity from a standpoint of featurewise matching, named by unsupervised dual hashing coding (UDC). It jointly learns the dual hashing codes on semantic tagging and sample content through factorizing a feature matching potential, which is allowed to bridge the semantic and heterogeneous gaps among different modalities simultaneously through maintaining the inter-modality-consistent semantic information and cross-modality-correlated sample content. In this way, each sample is uniquely coded by a head code on semanticwise tags, and tail codes on samplewise content. The dual coding design makes it very efficient for sample retrieval, in which the query sample only need to search for the retrieved ones with the same semantic tag, greatly narrowing down the search space. The proposed model avoids the calculation of massive sample-wise similarity and works with dual hashing coding scheme, which achieves a twofold efficiency enhancement for analyzing the large-scale and high-dimensional multimodal data. Extensive experiments have been conducted to demonstrate that it achieved superiority on computational time and retrieval performance.
- 11 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Emerging Topics in Computational Intelligence
- Vol. 8 (3) , 2529-2540
Unsupervised Feature Selection via Collaborative Embedding Learning
Unsupervised feature selection is vital in explanatory learning and remains challenging due to the difficulty of formulating a learnable model. Recently, graph embedding learning has gained widespread popularity in unsupervised learning, which extracts low-dimensional representation based on graph structure. Nevertheless, such an embedding scheme for unsupervised feature selection will distort original features due to the spatial transformation by extraction. To address this problem, this paper proposes a collaborative graph embedding model for unsupervised feature selection via jointly using soft-threshold and low-dimensional embedding learning. The former learns a threshold selection matrix for feature weighting in the original space. The latter extracts embedded representation in low-dimensional space to reveal the latent graph structure. By collaborative learning, the proposed method can simultaneously perform unsupervised feature selection in the original space and adaptive graph learning via dual embedding. Extensive experiments on five benchmark datasets demonstrate that the proposed method achieves superior performance compared to eight competing methods.
- 11 March 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Emerging Topics in Computational Intelligence
- Vol. 8 (3) , 2529-2540
Unsupervised Feature Selection via Collaborative Embedding Learning
Unsupervised feature selection is vital in explanatory learning and remains challenging due to the difficulty of formulating a learnable model. Recently, graph embedding learning has gained widespread popularity in unsupervised learning, which extracts low-dimensional representation based on graph structure. Nevertheless, such an embedding scheme for unsupervised feature selection will distort original features due to the spatial transformation by extraction. To address this problem, this paper proposes a collaborative graph embedding model for unsupervised feature selection via jointly using soft-threshold and low-dimensional embedding learning. The former learns a threshold selection matrix for feature weighting in the original space. The latter extracts embedded representation in low-dimensional space to reveal the latent graph structure. By collaborative learning, the proposed method can simultaneously perform unsupervised feature selection in the original space and adaptive graph learning via dual embedding. Extensive experiments on five benchmark datasets demonstrate that the proposed method achieves superior performance compared to eight competing methods.
- 29 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Emerging Topics in Computational Intelligence
- Vol. 8 (2) , 1841-1852
Collaborative Embedding Learning via Tensor Integration for Multi-View Clustering
Multi-view clustering exploits the complementary information of different views for comprehensive data analysis. Recently, graph learning techniques with low-dimensional embedding have been developed to learn consensus affinity graph for multi-view clustering. However, projecting data into the low-dimensional space has often resulted in the compression of data information, which is insufficient for graph learning. To address this challenge, this paper proposes a Collaborative Embedding Learning via Tensor (CELT) method, which learns intra-view affinity graphs for each view from both the original space and the low-dimensional space jointly. Additionally, all intra-view affinity graphs are stacked into a tensor, allowing the learning of a consensus affinity to capture inter-view consistency. In this way, an enhanced consensus affinity is obtained to improve the performance of multi-view clustering. Extensive experimental results on eight real-world datasets demonstrate that the proposed collaborative learning framework is effective for graph learning and outperforms competitive multi-view clustering methods.
- 29 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Emerging Topics in Computational Intelligence
- Vol. 8 (2) , 1841-1852
Collaborative Embedding Learning via Tensor Integration for Multi-View Clustering
Multi-view clustering exploits the complementary information of different views for comprehensive data analysis. Recently, graph learning techniques with low-dimensional embedding have been developed to learn consensus affinity graph for multi-view clustering. However, projecting data into the low-dimensional space has often resulted in the compression of data information, which is insufficient for graph learning. To address this challenge, this paper proposes a Collaborative Embedding Learning via Tensor (CELT) method, which learns intra-view affinity graphs for each view from both the original space and the low-dimensional space jointly. Additionally, all intra-view affinity graphs are stacked into a tensor, allowing the learning of a consensus affinity to capture inter-view consistency. In this way, an enhanced consensus affinity is obtained to improve the performance of multi-view clustering. Extensive experimental results on eight real-world datasets demonstrate that the proposed collaborative learning framework is effective for graph learning and outperforms competitive multi-view clustering methods.