Lishan Qiao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Lishan Qiao has a total of 66 co-authors. Here are the top co-authors by number of shared publications:
Limei Zhang
Shandong Jianzhu University
4 shared publications- Mingxia Liu
University of North Carolina at Chapel Hill
4 shared publications Dinggang Shen
School of Biomedical Engineering, ShanghaiTech University, Shanghai, China
3 shared publications- Xiaoming Xi
School of Computer Science and Technology, Shandong Jianzhu University, Jinan, China
3 shared publications Yanfang Xue
School of Mathematics Science, Liaocheng University, Liaocheng, China
3 shared publications- Chris Nugent
School of Computing, Ulster University, Belfast, UK
2 shared publications
Latest Publications
- 29 June 2026
- Published by Elsevier in Neural Networks
- Vol. 205  (Pt A) , 109305
AdapHBNA: Adaptive hierarchical spatio-temporal brain network analysis for brain disease detection
- 29 June 2026
- Published by Elsevier in Neural Networks
- Vol. 205  (Pt A) , 109305
AdapHBNA: Adaptive hierarchical spatio-temporal brain network analysis for brain disease detection
- Top Cited Papers
- 12 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. PP  (15209210) , 1-12
Difficulty-aware pseudo-label correction Network for Fine-Grained Classification of Choroidal Neovascularization in OCT Images
Choroidal neovascularization (CNV) classification is a fine-grained classification task. Accurate classification of CNV in optical coherence tomography (OCT) images is crucial for clinical treatment. However, image acquisition noise degrades image quality and exacerbates confirmation bias from class imbalance in medical datasets. Moreover, significant inter-class ambiguity in fine-grained categories can misclassify informative samples (e.g., hard samples or minority class samples) when generating pseudo-labels, leading to sub-optimal classifiers. To address these challenges, we propose a difficulty-aware pseudo-label correction network (DPLC-Net). Specifically, we designed a robust feature mining module using feature similarity loss to maintain consistency between generated adversarial and original samples, enabling noise-resistant feature learning. A difficulty-aware pseudo-label correction module mines and corrects potential noisy pseudo-labels to improve classification performance. Finally, to alleviate data bias and leverage all unlabeled samples, we integrated a hybrid consistency and pseudo-labeling module comprising adaptive weighted consistency loss (AWCL) and class-aware dynamic threshold strategy (CDTS). AWCL adaptively learns weights for unlabeled samples, effectively utilizing all unlabeled data through weighted consistency loss. CDTS dynamically adjusts confidence thresholds based on class distribution and model learning status, improving pseudo-label quantity and quality. Experiments on private and public OCT datasets demonstrate that our method outperforms state-of-the-art methods.
- 12 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. PP  (15209210) , 1-12
Top Cited PapersDifficulty-aware pseudo-label correction Network for Fine-Grained Classification of Choroidal Neovascularization in OCT Images
Choroidal neovascularization (CNV) classification is a fine-grained classification task. Accurate classification of CNV in optical coherence tomography (OCT) images is crucial for clinical treatment. However, image acquisition noise degrades image quality and exacerbates confirmation bias from class imbalance in medical datasets. Moreover, significant inter-class ambiguity in fine-grained categories can misclassify informative samples (e.g., hard samples or minority class samples) when generating pseudo-labels, leading to sub-optimal classifiers. To address these challenges, we propose a difficulty-aware pseudo-label correction network (DPLC-Net). Specifically, we designed a robust feature mining module using feature similarity loss to maintain consistency between generated adversarial and original samples, enabling noise-resistant feature learning. A difficulty-aware pseudo-label correction module mines and corrects potential noisy pseudo-labels to improve classification performance. Finally, to alleviate data bias and leverage all unlabeled samples, we integrated a hybrid consistency and pseudo-labeling module comprising adaptive weighted consistency loss (AWCL) and class-aware dynamic threshold strategy (CDTS). AWCL adaptively learns weights for unlabeled samples, effectively utilizing all unlabeled data through weighted consistency loss. CDTS dynamically adjusts confidence thresholds based on class distribution and model learning status, improving pseudo-label quantity and quality. Experiments on private and public OCT datasets demonstrate that our method outperforms state-of-the-art methods.
- 18 January 2026
- Published by Elsevier in Neural Networks
- Vol. 198, 108621
GIN-transformer based pairwise graph contrastive learning framework
- 18 January 2026
- Published by Elsevier in Neural Networks
- Vol. 198, 108621
GIN-transformer based pairwise graph contrastive learning framework
- 24 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 30  (7) , 6058-6071
Adaptive High-Order Fusion Learning for Brain Disorder Detection
The functional brain network (FBN) serves as an important tool for investigating neurological and mental disorders. Unlike many traditional networks whose structures are often known in advance, FBNs need to be estimated from neuroimaging or electrophysiological data, and their quality generally determines the performance of downstream tasks, particularly in disorder detection. Recent studies have shown that high-order FBNs tend to achieve better discriminative performance, while some other work indicates that increasing the order of FBN does not necessarily bring additional gains in discriminative performance and may even lead to a rapid decline in discriminability. To fully leverage information across different-order FBNs and identify optimal order for downstream tasks, we design an adaptive high-order FBN fusion learning framework (AHFL) with attention mechanism for brain disorder detection. Specifically, we first construct a series of FBNs with continuously increasing orders and propose a data-driven approach to evaluate each order's contribution to the classification performance. The self-attention mechanism is employed to capture contextual dependencies during the sequential generation of multi-order FBNs, thereby offering a natural fusion approach. Experimental evidence shows that the proposed method achieves superior performance compared to the baseline. In particular, we find that third-order FBN is achieved the highest weights, playing a crucial role for the detection of autism spectrum disorder (ASD), whereas second-order FBN is the most effective for identifying patients with major depressive disorder (MDD). Our findings advance the identification of discriminative high-order FBNs and establish a generalizable diagnostic framework for brain disorders.
- 24 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 30  (7) , 6058-6071
Adaptive High-Order Fusion Learning for Brain Disorder Detection
The functional brain network (FBN) serves as an important tool for investigating neurological and mental disorders. Unlike many traditional networks whose structures are often known in advance, FBNs need to be estimated from neuroimaging or electrophysiological data, and their quality generally determines the performance of downstream tasks, particularly in disorder detection. Recent studies have shown that high-order FBNs tend to achieve better discriminative performance, while some other work indicates that increasing the order of FBN does not necessarily bring additional gains in discriminative performance and may even lead to a rapid decline in discriminability. To fully leverage information across different-order FBNs and identify optimal order for downstream tasks, we design an adaptive high-order FBN fusion learning framework (AHFL) with attention mechanism for brain disorder detection. Specifically, we first construct a series of FBNs with continuously increasing orders and propose a data-driven approach to evaluate each order's contribution to the classification performance. The self-attention mechanism is employed to capture contextual dependencies during the sequential generation of multi-order FBNs, thereby offering a natural fusion approach. Experimental evidence shows that the proposed method achieves superior performance compared to the baseline. In particular, we find that third-order FBN is achieved the highest weights, playing a crucial role for the detection of autism spectrum disorder (ASD), whereas second-order FBN is the most effective for identifying patients with major depressive disorder (MDD). Our findings advance the identification of discriminative high-order FBNs and establish a generalizable diagnostic framework for brain disorders.
- 16 November 2025
- Published by Springer Nature
- p. 211-224
BrainGACCL: Brain Graph Adaptive Co-contrastive Learning with Universum Samples for fMRI-Based Brain Disease Detection
- 16 November 2025
- Published by Springer Nature
- p. 211-224
BrainGACCL: Brain Graph Adaptive Co-contrastive Learning with Universum Samples for fMRI-Based Brain Disease Detection
- Open Access
- 26 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 169866-169880
Research on Personalized Recommendation Based on Matrix Factorization, Clustering, and Deep Reinforcement Learning
To address the challenges of data sparsity, cold start, and insufficient dynamic adaptability in traditional recommendation systems, this paper proposes a personalized recommendation model named CDRL-MF, which integrates Matrix Factorization, multi-viewclustering, and Deep Reinforcement Learning. The method achieves performance improvement through a three-stage collaborative optimization process: First, a dual-channel architecture employing SingularValue Decomposition and an Artificial Neural Network generates high-quality user and item embeddings. Second, multi-view K-means clustering is introduced to construct precise user interest clusters by synthesizing user rating patterns, statistical attributes, and content features. Finally, a Cluster-Guided Deep Reinforcement Recommendation framework is designed, where a DDPG-based agent integrates user states, item cluster features, and real-time feedback to achieve continuous dynamic optimization of recommendation policies. Experimental results on the MovieLens 1M dataset demonstrate that the CDRL-MF model significantly outperforms multiple baseline models across key evaluation metrics, including rating prediction (MAE, RMSE) and Top-N recommendation (Precision, Recall, F1, NDCG). Furthermore, the model exhibits excellent balancing capabilities in recommendation diversity, novelty, and user group fairness. By incorporating differential privacy and federated learning mechanisms, it maintains acceptable performance trade-offs while ensuring user privacy protection. Additional experiments on large-scale datasets such as Amazon Reviews and Netflix Prize further validate its robust generalization capability and practicality.
- 26 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 169866-169880
Open AccessResearch on Personalized Recommendation Based on Matrix Factorization, Clustering, and Deep Reinforcement Learning
To address the challenges of data sparsity, cold start, and insufficient dynamic adaptability in traditional recommendation systems, this paper proposes a personalized recommendation model named CDRL-MF, which integrates Matrix Factorization, multi-viewclustering, and Deep Reinforcement Learning. The method achieves performance improvement through a three-stage collaborative optimization process: First, a dual-channel architecture employing SingularValue Decomposition and an Artificial Neural Network generates high-quality user and item embeddings. Second, multi-view K-means clustering is introduced to construct precise user interest clusters by synthesizing user rating patterns, statistical attributes, and content features. Finally, a Cluster-Guided Deep Reinforcement Recommendation framework is designed, where a DDPG-based agent integrates user states, item cluster features, and real-time feedback to achieve continuous dynamic optimization of recommendation policies. Experimental results on the MovieLens 1M dataset demonstrate that the CDRL-MF model significantly outperforms multiple baseline models across key evaluation metrics, including rating prediction (MAE, RMSE) and Top-N recommendation (Precision, Recall, F1, NDCG). Furthermore, the model exhibits excellent balancing capabilities in recommendation diversity, novelty, and user group fairness. By incorporating differential privacy and federated learning mechanisms, it maintains acceptable performance trade-offs while ensuring user privacy protection. Additional experiments on large-scale datasets such as Amazon Reviews and Netflix Prize further validate its robust generalization capability and practicality.
- 5 September 2025
- Published by Elsevier in Neural Networks
- Vol. 193, 108081
GoFormer: A GoLPP inspired transformer for functional brain graph learning and classification
- 5 September 2025
- Published by Elsevier in Neural Networks
- Vol. 193, 108081
GoFormer: A GoLPP inspired transformer for functional brain graph learning and classification
- 20 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 36  (2) , 1601-1613
Prior Distribution Guided Gaussian Mixture Variational Autoencoder (PDGM-VAE) for Image Generation
Variational Autoencoder(VAE) combines the ideas of autoencoders and variational inference, introducing the concept of latent space and variational inference to endow autoencoders to generate new images. VAE typically assumes that data follows a Gaussian distribution, but real data may follow other distributions. This inconsistency between the assumption and the true distribution can affect the modeling and reconstruction capabilities of VAE, which makes it difficult for traditional models to accurately capture the true distribution. To address the aforementioned issues, we propose a Prior Distribution Guided Gaussian Mixture Variational Autoencoder(PDGM-VAE). Specifically, we construct a Gaussian Mixture Prior Learner (GMPL) to capture complex features of the data distribution, enabling the model to learn and obtain a Gaussian mixture distribution that is reasonable and close to the real data distributions, which is then used as the prior distribution in the network. Furthermore, we build a Semantic-Aware Module with Embedded Prior Distribution (SAMEPD), integrating data and label information to learn the distribution parameters, enabling the network to learn and utilize the semantic knowledge contained in the labels. During training, by approximating the posterior distribution to the prior distribution, we enhance the model’s modeling and reconstruction capabilities, improving the quality of generated images. We evaluated the image generation task on five public datasets, and based on the FID metric, our proposed method outperformed other VAE methods.
- 20 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 36  (2) , 1601-1613
Prior Distribution Guided Gaussian Mixture Variational Autoencoder (PDGM-VAE) for Image Generation
Variational Autoencoder(VAE) combines the ideas of autoencoders and variational inference, introducing the concept of latent space and variational inference to endow autoencoders to generate new images. VAE typically assumes that data follows a Gaussian distribution, but real data may follow other distributions. This inconsistency between the assumption and the true distribution can affect the modeling and reconstruction capabilities of VAE, which makes it difficult for traditional models to accurately capture the true distribution. To address the aforementioned issues, we propose a Prior Distribution Guided Gaussian Mixture Variational Autoencoder(PDGM-VAE). Specifically, we construct a Gaussian Mixture Prior Learner (GMPL) to capture complex features of the data distribution, enabling the model to learn and obtain a Gaussian mixture distribution that is reasonable and close to the real data distributions, which is then used as the prior distribution in the network. Furthermore, we build a Semantic-Aware Module with Embedded Prior Distribution (SAMEPD), integrating data and label information to learn the distribution parameters, enabling the network to learn and utilize the semantic knowledge contained in the labels. During training, by approximating the posterior distribution to the prior distribution, we enhance the model’s modeling and reconstruction capabilities, improving the quality of generated images. We evaluated the image generation task on five public datasets, and based on the FID metric, our proposed method outperformed other VAE methods.
- 22 July 2025
- Published by Elsevier in Neural Networks
- Vol. 192, 107891
GSAformer: Group sparse attention transformer for functional brain network analysis
- 22 July 2025
- Published by Elsevier in Neural Networks
- Vol. 192, 107891
GSAformer: Group sparse attention transformer for functional brain network analysis
- 21 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 29  (11) , 8460-8473
Miniformer: A Minimalist Transformer for Brain Functional Networks Analysis
Learning to estimate and classify brain functional networks (BFNs) has become an increasingly important way of predicting neurological or mental disorders at their early stages. The traditional methods conduct BFN estimation and classification in two separate steps, thus preventing the interaction and joint optimization. In contrast, Transformer provides a natural architecture to learn BFNs with downstream tasks in an end-to-end manner. Despite their great potential, Transformer-based methods involve a large number of parameters that need to be learnt from big data and often lead to poor model interpretability. Considering the challenge in acquiring data and the high demand for model interpretability in medical scenarios, in this paper, we propose a minimalist Transformer architecture, referred to as Miniformer, by simplifying the projection matrices in the self-attention module into a single diagonal matrix, which greatly reduces the number of parameters, alleviates the risk of overfitting, and improves the interpretability. Additionally, the clear physical meaning of parameters in Miniformer makes the integration of domain knowledge or prior easier and more natural. Therefore, we further develop two variants of Miniformer by incorporating sparsity for removing potentially noisy time points from fMRI signals, and smoothness for capturing the temporal correlations in fMRI signals, respectively. To evaluate the effectiveness of the proposed methods, we perform brain disease diagnosis experiments on three public datasets. The results show that Miniformer and its variants tend to achieve higher classification performance than comparison methods with good interpretability.
- 21 May 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal of Biomedical and Health Informatics
- Vol. 29  (11) , 8460-8473
Miniformer: A Minimalist Transformer for Brain Functional Networks Analysis
Learning to estimate and classify brain functional networks (BFNs) has become an increasingly important way of predicting neurological or mental disorders at their early stages. The traditional methods conduct BFN estimation and classification in two separate steps, thus preventing the interaction and joint optimization. In contrast, Transformer provides a natural architecture to learn BFNs with downstream tasks in an end-to-end manner. Despite their great potential, Transformer-based methods involve a large number of parameters that need to be learnt from big data and often lead to poor model interpretability. Considering the challenge in acquiring data and the high demand for model interpretability in medical scenarios, in this paper, we propose a minimalist Transformer architecture, referred to as Miniformer, by simplifying the projection matrices in the self-attention module into a single diagonal matrix, which greatly reduces the number of parameters, alleviates the risk of overfitting, and improves the interpretability. Additionally, the clear physical meaning of parameters in Miniformer makes the integration of domain knowledge or prior easier and more natural. Therefore, we further develop two variants of Miniformer by incorporating sparsity for removing potentially noisy time points from fMRI signals, and smoothness for capturing the temporal correlations in fMRI signals, respectively. To evaluate the effectiveness of the proposed methods, we perform brain disease diagnosis experiments on three public datasets. The results show that Miniformer and its variants tend to achieve higher classification performance than comparison methods with good interpretability.