Mujie Liu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Mujie Liu has a total of 47 co-authors. Here are the top co-authors by number of shared publications:
- Feng Xia
School of Computing Technologies, RMIT University
13 shared publications - Jing Ren
RMIT University, Australia
7 shared publications - Falih Gozi Febrinanto
Institute of Innovation, Science and Sustainability, Federation University Australia, Ballarat, VIC, Australia
5 shared publications - Chenze Wang
Zhejiang Gongshang University
4 shared publications - Liping Chen
RMIT University
4 shared publications - Nguyen Linh Dan Le
RMIT University, Australia
4 shared publications
Latest Publications
- Open Access
- 16 November 2026
- Published by Association for Computing Machinery (ACM) in ACM Computing Surveys
Multimodal Healthcare Foundation Models: Techniques, Challenges, and Opportunities
With the rapid advance of multimodal foundation models, multimodal healthcare foundation models (MHFMs) are increasingly being adopted for reliable, intelligent clinical applications. They promise to advance precision medicine, streamline clinical workflows, and ultimately improve patient outcomes. Yet the healthcare domain imposes stringent requirements, including accuracy, safety, transparency, and trustworthiness, that demand techniques beyond those used for general-purpose multimodal foundation models. However, existing surveys largely take application-centric or data-oriented views and often underplay the technical underpinnings and domain-specific adaptations. To fill this gap, this survey explores and summarizes MHFMs from a technical perspective. Specifically, we provide a systematical taxonomy of MHFMs based on their methodologies, summarizing multimodal foundation model techniques that are applicable to the healthcare domain. In addition, we synthesize the key challenges and opportunities for deploying MHFMs in real-world healthcare settings, providing a roadmap for future research.
- 16 November 2026
- Published by Association for Computing Machinery (ACM) in ACM Computing Surveys
Open AccessMultimodal Healthcare Foundation Models: Techniques, Challenges, and Opportunities
With the rapid advance of multimodal foundation models, multimodal healthcare foundation models (MHFMs) are increasingly being adopted for reliable, intelligent clinical applications. They promise to advance precision medicine, streamline clinical workflows, and ultimately improve patient outcomes. Yet the healthcare domain imposes stringent requirements, including accuracy, safety, transparency, and trustworthiness, that demand techniques beyond those used for general-purpose multimodal foundation models. However, existing surveys largely take application-centric or data-oriented views and often underplay the technical underpinnings and domain-specific adaptations. To fill this gap, this survey explores and summarizes MHFMs from a technical perspective. Specifically, we provide a systematical taxonomy of MHFMs based on their methodologies, summarizing multimodal foundation model techniques that are applicable to the healthcare domain. In addition, we synthesize the key challenges and opportunities for deploying MHFMs in real-world healthcare settings, providing a roadmap for future research.
- Open Access
- 21 October 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Computing for Healthcare
Explainable Graph Neural Networks: Understanding Brain Connectivity and Biomarkers in Dementia
Dementia is a progressive neurodegenerative disorder with multiple etiologies, including Alzheimer's disease, Parkinson's disease, frontotemporal dementia, and vascular dementia. Its clinical and biological heterogeneity makes diagnosis and subtype differentiation highly challenging. Graph Neural Networks (GNNs) have recently shown strong potential in modeling brain connectivity, but their limited robustness, data scarcity, and lack of interpretability constrain clinical adoption. Explainable Graph Neural Networks (XGNNs) have emerged to address these barriers by combining graph-based learning with interpretability, enabling the identification of disease-relevant biomarkers, analysis of brain network disruptions, and provision of transparent insights for clinicians. This paper presents the first comprehensive review dedicated to XGNNs in dementia research. We examine their applications across Alzheimer's disease, Parkinson's disease, mild cognitive impairment, and multi-disease diagnosis. A taxonomy of explainability methods tailored for dementia-related tasks is introduced, alongside comparisons of existing models in clinical scenarios. We also highlight challenges such as limited generalizability, underexplored domains, and the integration of Large Language Models (LLMs) for early detection. By outlining both progress and open problems, this review aims to guide future work toward trustworthy, clinically meaningful, and scalable use of XGNNs in dementia research.
- 21 October 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Computing for Healthcare
Open AccessExplainable Graph Neural Networks: Understanding Brain Connectivity and Biomarkers in Dementia
Dementia is a progressive neurodegenerative disorder with multiple etiologies, including Alzheimer's disease, Parkinson's disease, frontotemporal dementia, and vascular dementia. Its clinical and biological heterogeneity makes diagnosis and subtype differentiation highly challenging. Graph Neural Networks (GNNs) have recently shown strong potential in modeling brain connectivity, but their limited robustness, data scarcity, and lack of interpretability constrain clinical adoption. Explainable Graph Neural Networks (XGNNs) have emerged to address these barriers by combining graph-based learning with interpretability, enabling the identification of disease-relevant biomarkers, analysis of brain network disruptions, and provision of transparent insights for clinicians. This paper presents the first comprehensive review dedicated to XGNNs in dementia research. We examine their applications across Alzheimer's disease, Parkinson's disease, mild cognitive impairment, and multi-disease diagnosis. A taxonomy of explainability methods tailored for dementia-related tasks is introduced, alongside comparisons of existing models in clinical scenarios. We also highlight challenges such as limited generalizability, underexplored domains, and the integration of Large Language Models (LLMs) for early detection. By outlining both progress and open problems, this review aims to guide future work toward trustworthy, clinically meaningful, and scalable use of XGNNs in dementia research.
- Open Access
- 24 June 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Intelligent Systems and Technology
- Vol. 17 (5) , 1-25
Multi-Scale Diffusion for Bio-topological Representation Learning on Multimodal Brain Graphs
Understanding how neurological disorders disrupt brain network topology requires modeling the brain's hierarchical and multiscale organization. However, existing brain graph learning methods predominantly focus on single-scale structures, limiting their ability to capture interdependent topological alterations across modalities and scales. We propose BrainSTORE, a multi-scale diffusion framework for bio-topological representation learning on multimodal brain graphs that jointly models structural and functional connectivity. BrainSTORE introduces a multiscale community-detection strategy that constructs hierarchically consistent topological priors across local, meso-, and global scales while preserving cross-modal dependencies. These priors are integrated into the diffusion process via a scale-specific noise-scheduling mechanism that guides perturbations along biologically meaningful structural pathways. A unified joint denoising architecture further enforces cross-modal consistency, enabling the extraction of both modality-specific and shared bio-topological features. Extensive experiments on two real-world neuroimaging datasets demonstrate that BrainSTORE consistently outperforms state-of-the-art methods in brain disease detection. Beyond improved predictive performance, BrainSTORE provides a principled framework for investigating scale-dependent and modality-coupled topological alterations associated with neurological disorders.
- 24 June 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Intelligent Systems and Technology
- Vol. 17 (5) , 1-25
Open AccessMulti-Scale Diffusion for Bio-topological Representation Learning on Multimodal Brain Graphs
Understanding how neurological disorders disrupt brain network topology requires modeling the brain's hierarchical and multiscale organization. However, existing brain graph learning methods predominantly focus on single-scale structures, limiting their ability to capture interdependent topological alterations across modalities and scales. We propose BrainSTORE, a multi-scale diffusion framework for bio-topological representation learning on multimodal brain graphs that jointly models structural and functional connectivity. BrainSTORE introduces a multiscale community-detection strategy that constructs hierarchically consistent topological priors across local, meso-, and global scales while preserving cross-modal dependencies. These priors are integrated into the diffusion process via a scale-specific noise-scheduling mechanism that guides perturbations along biologically meaningful structural pathways. A unified joint denoising architecture further enforces cross-modal consistency, enabling the extraction of both modality-specific and shared bio-topological features. Extensive experiments on two real-world neuroimaging datasets demonstrate that BrainSTORE consistently outperforms state-of-the-art methods in brain disease detection. Beyond improved predictive performance, BrainSTORE provides a principled framework for investigating scale-dependent and modality-coupled topological alterations associated with neurological disorders.
- 26 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cognitive and Developmental Systems
- Vol. PP (23798920) , 1-13
Spider-Brain: Spike-inspired Graph Learning of Asynchronous and Hierarchical Brain Dynamics
Current methods for analyzing dynamic brain connectivity assume global synchrony, overlooking the inherent asynchrony and hierarchical nature of neural processes, thereby leading to incomplete or biased interpretations of brain dynamics. To address these limitations, we propose Spider-Brain, a spike-inspired dynamic graph structural learning (GSL) framework that models the brain as an evolving system with complex temporal organization. Spider-Brain incorporates a spike-inspired selective activation mechanism in which node updates occur through region-specific, event-triggered propagation, effectively capturing heterogeneous temporal responses across brain regions. By combining a multiscale temporal sampling strategy with a selective state-space model (SSM), the framework encodes both short-term neural fluctuations and long-range temporal dependencies, enabling explicit modeling of hierarchical temporal structure. Experiments on two real-world psychiatric datasets show that Spider-Brain consistently outperforms state-of-the-art baselines in disorder classification and provides interpretable activation patterns. The framework also uncovers disorder-specific connectivity and regional activation signatures that align with established clinical and neuroscientific findings. This demonstrates its ability to identify interpretable patterns of brain connectivity and temporal activity. Our code is available at https://github.com/mjliu99/Spider-Brain.
- 26 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cognitive and Developmental Systems
- Vol. PP (23798920) , 1-13
Spider-Brain: Spike-inspired Graph Learning of Asynchronous and Hierarchical Brain Dynamics
Current methods for analyzing dynamic brain connectivity assume global synchrony, overlooking the inherent asynchrony and hierarchical nature of neural processes, thereby leading to incomplete or biased interpretations of brain dynamics. To address these limitations, we propose Spider-Brain, a spike-inspired dynamic graph structural learning (GSL) framework that models the brain as an evolving system with complex temporal organization. Spider-Brain incorporates a spike-inspired selective activation mechanism in which node updates occur through region-specific, event-triggered propagation, effectively capturing heterogeneous temporal responses across brain regions. By combining a multiscale temporal sampling strategy with a selective state-space model (SSM), the framework encodes both short-term neural fluctuations and long-range temporal dependencies, enabling explicit modeling of hierarchical temporal structure. Experiments on two real-world psychiatric datasets show that Spider-Brain consistently outperforms state-of-the-art baselines in disorder classification and provides interpretable activation patterns. The framework also uncovers disorder-specific connectivity and regional activation signatures that align with established clinical and neuroscientific findings. This demonstrates its ability to identify interpretable patterns of brain connectivity and temporal activity. Our code is available at https://github.com/mjliu99/Spider-Brain.
- Open Access
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 9593-9603
FairFRL: Fairness-aware Federated Representation Learning for Cross-domain Sequential Recommendation
- 12 April 2026
- Published by Association for Computing Machinery (ACM)
- p. 9593-9603
Open AccessFairFRL: Fairness-aware Federated Representation Learning for Cross-domain Sequential Recommendation
- Top Cited Papers
- 26 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Social Systems
- Vol. 13 (3) , 4131-4143
Causal Prompting for Implicit Sentiment Analysis With Large Language Models
Implicit sentiment analysis (ISA) aims to infer sentiment that is implied rather than explicitly stated, requiring models to perform deeper reasoning over subtle contextual cues. While recent prompting-based methods using large language models (LLMs) have shown promise in ISA, they often rely on majority voting over chain-of-thought (CoT) reasoning paths without evaluating their causal validity, making them susceptible to internal biases and spurious correlations. To address this challenge, we propose CAPITAL, a causal prompting framework that incorporates front-door adjustment into CoT reasoning. CAPITAL decomposes the overall causal effect into two components: the influence of the input prompt on the reasoning chains, and the impact of those chains on the final output. These components are estimated using encoder-based clustering and the NWGM approximation, with a contrastive learning objective used to better align the encoder’s representation with the LLM’s reasoning space. Experiments on benchmark ISA datasets with three LLMs demonstrate that CAPITAL consistently outperforms strong prompting baselines in both accuracy and robustness, particularly under adversarial conditions. This work offers a principled approach to integrating causal inference into LLM prompting and highlights its benefits for bias-aware sentiment reasoning. The source code and case study are available at: https://github.com/whZ62/CAPITAL.
- 26 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Social Systems
- Vol. 13 (3) , 4131-4143
Top Cited PapersCausal Prompting for Implicit Sentiment Analysis With Large Language Models
Implicit sentiment analysis (ISA) aims to infer sentiment that is implied rather than explicitly stated, requiring models to perform deeper reasoning over subtle contextual cues. While recent prompting-based methods using large language models (LLMs) have shown promise in ISA, they often rely on majority voting over chain-of-thought (CoT) reasoning paths without evaluating their causal validity, making them susceptible to internal biases and spurious correlations. To address this challenge, we propose CAPITAL, a causal prompting framework that incorporates front-door adjustment into CoT reasoning. CAPITAL decomposes the overall causal effect into two components: the influence of the input prompt on the reasoning chains, and the impact of those chains on the final output. These components are estimated using encoder-based clustering and the NWGM approximation, with a contrastive learning objective used to better align the encoder’s representation with the LLM’s reasoning space. Experiments on benchmark ISA datasets with three LLMs demonstrate that CAPITAL consistently outperforms strong prompting baselines in both accuracy and robustness, particularly under adversarial conditions. This work offers a principled approach to integrating causal inference into LLM prompting and highlights its benefits for bias-aware sentiment reasoning. The source code and case study are available at: https://github.com/whZ62/CAPITAL.
- 27 November 2025
- Published by Springer Nature
- p. 359-372
Structure Matters: Brain Graph Augmentation via Learnable Edge Masking for Data-Efficient Psychiatric Diagnosis
- 27 November 2025
- Published by Springer Nature
- p. 359-372
Structure Matters: Brain Graph Augmentation via Learnable Edge Masking for Data-Efficient Psychiatric Diagnosis
- 27 November 2025
- Published by Springer Nature
- p. 347-358
Personalized Federated Graph Learning for Heterogeneous Incomplete EHRs
- 27 November 2025
- Published by Springer Nature
- p. 347-358
Personalized Federated Graph Learning for Heterogeneous Incomplete EHRs
- Open Access
Explainable Graph Neural Networks: Understanding Brain Connectivity and Biomarkers in Dementia
Dementia is a progressive neurodegenerative disorder with multiple etiologies, including Alzheimer's disease, Parkinson's disease, frontotemporal dementia, and vascular dementia. Its clinical and biological heterogeneity makes diagnosis and subtype differentiation highly challenging. Graph Neural Networks (GNNs) have recently shown strong potential in modeling brain connectivity, but their limited robustness, data scarcity, and lack of interpretability constrain clinical adoption. Explainable Graph Neural Networks (XGNNs) have emerged to address these barriers by combining graph-based learning with interpretability, enabling the identification of disease-relevant biomarkers, analysis of brain network disruptions, and provision of transparent insights for clinicians. This paper presents the first comprehensive review dedicated to XGNNs in dementia research. We examine their applications across Alzheimer's disease, Parkinson's disease, mild cognitive impairment, and multi-disease diagnosis. A taxonomy of explainability methods tailored for dementia-related tasks is introduced, alongside comparisons of existing models in clinical scenarios. We also highlight challenges such as limited generalizability, underexplored domains, and the integration of Large Language Models (LLMs) for early detection. By outlining both progress and open problems, this review aims to guide future work toward trustworthy, clinically meaningful, and scalable use of XGNNs in dementia research.
Open AccessExplainable Graph Neural Networks: Understanding Brain Connectivity and Biomarkers in Dementia
Dementia is a progressive neurodegenerative disorder with multiple etiologies, including Alzheimer's disease, Parkinson's disease, frontotemporal dementia, and vascular dementia. Its clinical and biological heterogeneity makes diagnosis and subtype differentiation highly challenging. Graph Neural Networks (GNNs) have recently shown strong potential in modeling brain connectivity, but their limited robustness, data scarcity, and lack of interpretability constrain clinical adoption. Explainable Graph Neural Networks (XGNNs) have emerged to address these barriers by combining graph-based learning with interpretability, enabling the identification of disease-relevant biomarkers, analysis of brain network disruptions, and provision of transparent insights for clinicians. This paper presents the first comprehensive review dedicated to XGNNs in dementia research. We examine their applications across Alzheimer's disease, Parkinson's disease, mild cognitive impairment, and multi-disease diagnosis. A taxonomy of explainability methods tailored for dementia-related tasks is introduced, alongside comparisons of existing models in clinical scenarios. We also highlight challenges such as limited generalizability, underexplored domains, and the integration of Large Language Models (LLMs) for early detection. By outlining both progress and open problems, this review aims to guide future work toward trustworthy, clinically meaningful, and scalable use of XGNNs in dementia research.
- Open Access
- 9 September 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Computing for Healthcare
Data-Efficient Psychiatric Disorder Detection via Self-supervised Learning on Frequency-enhanced Brain Networks
Psychiatric disorders involve complex neural activity changes, with functional magnetic resonance imaging (fMRI) data serving as key diagnostic evidence. However, data scarcity and the diverse nature of fMRI information pose significant challenges. While graph-based self-supervised learning (SSL) methods have shown promise in brain network analysis, they primarily focus on time-domain representations, often overlooking the rich information embedded in the frequency domain. To overcome these limitations, we propose F requency- E nhanced Net work (FENet), a novel SSL framework specially designed for fMRI data that integrates time-domain and frequency-domain information to improve psychiatric disorder detection in small-sample datasets. FENet constructs multi-view brain networks based on the inherent properties of fMRI data, explicitly incorporating frequency information into the learning process of representation. Additionally, it employs domain-specific encoders to capture temporal-spectral characteristics, including an efficient frequency-domain encoder that highlights disease-relevant frequency features. Finally, FENet introduces a domain consistency-guided learning objective, which balances the utilization of diverse information and generates frequency-enhanced brain graph representations. Experiments on two real-world medical datasets demonstrate that FENet outperforms state-of-the-art methods while maintaining strong performance in minimal data conditions. Furthermore, we analyze the correlation between various frequency-domain features and psychiatric disorders, emphasizing the critical role of high-frequency information in disorder detection.
- 9 September 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Computing for Healthcare
Open AccessData-Efficient Psychiatric Disorder Detection via Self-supervised Learning on Frequency-enhanced Brain Networks
Psychiatric disorders involve complex neural activity changes, with functional magnetic resonance imaging (fMRI) data serving as key diagnostic evidence. However, data scarcity and the diverse nature of fMRI information pose significant challenges. While graph-based self-supervised learning (SSL) methods have shown promise in brain network analysis, they primarily focus on time-domain representations, often overlooking the rich information embedded in the frequency domain. To overcome these limitations, we propose F requency- E nhanced Net work (FENet), a novel SSL framework specially designed for fMRI data that integrates time-domain and frequency-domain information to improve psychiatric disorder detection in small-sample datasets. FENet constructs multi-view brain networks based on the inherent properties of fMRI data, explicitly incorporating frequency information into the learning process of representation. Additionally, it employs domain-specific encoders to capture temporal-spectral characteristics, including an efficient frequency-domain encoder that highlights disease-relevant frequency features. Finally, FENet introduces a domain consistency-guided learning objective, which balances the utilization of diverse information and generates frequency-enhanced brain graph representations. Experiments on two real-world medical datasets demonstrate that FENet outperforms state-of-the-art methods while maintaining strong performance in minimal data conditions. Furthermore, we analyze the correlation between various frequency-domain features and psychiatric disorders, emphasizing the critical role of high-frequency information in disorder detection.