Liping Mu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Liping Mu has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Guangyu Huo
School of Information Science and Technology (School of Artificial Intelligence), Beijing Forestry University, Beijing, China
2 shared publications - Zhibo Chen
School of Information Science and Technology, Beijing Forestry University, Beijing, China
2 shared publications Rui Zhang
School of Information Science and Technology (School of Artificial Intelligence), Beijing Forestry University, Beijing, China
1 shared publication- Rui Zhang1 shared publication
Wenda Luo
School of Information Science and Technology (School of Artificial Intelligence), Beijing Forestry University, Beijing, China
1 shared publication- Wenda Luo1 shared publication
Latest Publications
- 24 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 64 (01962892) , 1-15
PSTH-CDPM: A Physics-Guided Spatiotemporal Hypergraph Conditional Denoising Diffusion Probabilistic Model for Invasive Alien Species Distribution Prediction
The impact of invasive alien species on ecosystems and economies is increasingly severe, making accurate predictions of their distribution crucial for effective management strategies. In response to this demand, this study proposes a species distribution prediction approach named Physics-Guided Spatio-Temporal Hypergraph Conditional Denoising Diffusion Probabilistic Model (PSTH-CDPM). To address the limitations of traditional approaches in modeling the physical evolution of environmental factors, this study proposes a Thermal Conduction-based Temperature Driven Module, which leverages heat diffusion convolution inspired by physical heat conduction to simulate the spatial propagation of environmental factors. A Spatio-Temporal Awareness Fusion Module is further designed to effectively integrate historical species distribution sequences with multi-modal environmental information, enabling joint modeling of species and environmental dynamics and enhancing the capture of complex spatiotemporal dependencies. To tackle the underutilization of multi-modal conditional information, this work pioneers the application of a diffusion model and employs the lightweight Convolutional Block Attention Module and a hierarchical Hypergraph Attention to enhance its denoising network. This design strengthens key feature representation and enables higher-order correlation modeling of cross-level multi-scale features. Two high-quality multi-temporal remote sensing–derived datasets are constructed. The experiment results indicate that the proposed method attains values of 0.9546, 0.9153, and 0.955 for the three key metrics—the kappa coefficient, the Intersection over Union, and the F1-score, respectively—demonstrating its high accuracy and robustness in spatial prediction tasks.
- 24 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 64 (01962892) , 1-15
PSTH-CDPM: A Physics-Guided Spatiotemporal Hypergraph Conditional Denoising Diffusion Probabilistic Model for Invasive Alien Species Distribution Prediction
The impact of invasive alien species on ecosystems and economies is increasingly severe, making accurate predictions of their distribution crucial for effective management strategies. In response to this demand, this study proposes a species distribution prediction approach named Physics-Guided Spatio-Temporal Hypergraph Conditional Denoising Diffusion Probabilistic Model (PSTH-CDPM). To address the limitations of traditional approaches in modeling the physical evolution of environmental factors, this study proposes a Thermal Conduction-based Temperature Driven Module, which leverages heat diffusion convolution inspired by physical heat conduction to simulate the spatial propagation of environmental factors. A Spatio-Temporal Awareness Fusion Module is further designed to effectively integrate historical species distribution sequences with multi-modal environmental information, enabling joint modeling of species and environmental dynamics and enhancing the capture of complex spatiotemporal dependencies. To tackle the underutilization of multi-modal conditional information, this work pioneers the application of a diffusion model and employs the lightweight Convolutional Block Attention Module and a hierarchical Hypergraph Attention to enhance its denoising network. This design strengthens key feature representation and enables higher-order correlation modeling of cross-level multi-scale features. Two high-quality multi-temporal remote sensing–derived datasets are constructed. The experiment results indicate that the proposed method attains values of 0.9546, 0.9153, and 0.955 for the three key metrics—the kappa coefficient, the Intersection over Union, and the F1-score, respectively—demonstrating its high accuracy and robustness in spatial prediction tasks.
- Open Access
- 27 November 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 302
InvaderDefender: Multimodal recognition of invasive alien species via a large language model with vision-guided targeted RAG
- 27 November 2025
- Published by Elsevier in Expert Systems with Applications
- Vol. 302
Open AccessInvaderDefender: Multimodal recognition of invasive alien species via a large language model with vision-guided targeted RAG