Yanli Wan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Yanli Wan has a total of 35 co-authors. Here are the top co-authors by number of shared publications:
- Hongpu Hu
Institute of Medical Information/Medical Library, Chinese Academy of Medical Sciences & Peking Union Medical College, China
3 shared publications - Yingshuai Wang
Institute of Medical Information/Library, Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing, China
3 shared publications Chongyang Zhang
Institute of Medical Information/Library, Chinese Academy of Medical Sciences and Peking Union Medical College
2 shared publications- Guoqiang Sun
Information Center, Chinese Academy of Medical Sciences, Peking Union Medical College Hospital, Beijing, China
2 shared publications Qiyan Yan
School of Public Administration, Southwest Minzu University, Chengdu, Sichuan, China
2 shared publicationsShanshan Zhao
Institute of Medical Information/Library, Chinese Academy of Medical Sciences and Peking Union Medical College
2 shared publications
Latest Publications
- Open Access
- 10 April 2026
- Published by Frontiers Media SA in Frontiers in Aging Neuroscience
- Vol. 18, 1657578
Research on Alzheimer’s disease MRI image classification based on spatial attention mechanism
Introduction: Early diagnosis of Alzheimer’s Disease (AD) is crucial for improving patient quality of life and treatment outcomes. However, accurately classifying MRI scans of AD remains challenging due to the subtle and spatially complex nature of lesion regions. This study proposes a novel bidirectional spatial attention mechanism to enhance the focus on key pathological features in AD MRI images, aiming to improve classification accuracy and support earlier intervention. Methods: To enhance model performance, we introduced a customized bidirectional spatial attention module (ATT) integrated into a Swin-Tiny Transformer backbone. Unlike conventional attention methods, the ATT module generates spatial attention maps by adaptively pooling features along both vertical and horizontal orientations, allowing refined adjustment of attention weights across different image regions. Furthermore, to address issues of limited sample size and class imbalance, we employed data augmentation and expansion strategies, enriching the diversity of training data. The model was trained and evaluated on the augmented OASIS1 dataset. Results: The improved Swin-Tiny+ATT model demonstrated significant performance gains across all key metrics on the augmented dataset. Compared to the baseline Swin Transformer, accuracy improved from 84.83% to 87.96%, recall from 89.82% to 91.92%, precision from 85.27% to 91.98%, and the F1 score from 87.26% to 91.89%. These results confirm that the ATT module effectively enhances the model’s ability to capture complex spatial features and identify critical lesion regions. Discussion: The proposed Swin-Tiny+ATT model exhibits strong potential for improving MRI-based classification of Alzheimer’s Disease. The bidirectional spatial attention mechanism successfully directs the model’s focus to relevant anatomical regions, contributing to higher precision and recall. Combined with data augmentation strategies, the approach mitigates class imbalance and enhances generalization. This work provides a promising deep learning framework to support early and accurate diagnosis of AD, with implications for clinical decision-making and personalized treatment planning.
- 10 April 2026
- Published by Frontiers Media SA in Frontiers in Aging Neuroscience
- Vol. 18, 1657578
Open AccessResearch on Alzheimer’s disease MRI image classification based on spatial attention mechanism
Introduction: Early diagnosis of Alzheimer’s Disease (AD) is crucial for improving patient quality of life and treatment outcomes. However, accurately classifying MRI scans of AD remains challenging due to the subtle and spatially complex nature of lesion regions. This study proposes a novel bidirectional spatial attention mechanism to enhance the focus on key pathological features in AD MRI images, aiming to improve classification accuracy and support earlier intervention. Methods: To enhance model performance, we introduced a customized bidirectional spatial attention module (ATT) integrated into a Swin-Tiny Transformer backbone. Unlike conventional attention methods, the ATT module generates spatial attention maps by adaptively pooling features along both vertical and horizontal orientations, allowing refined adjustment of attention weights across different image regions. Furthermore, to address issues of limited sample size and class imbalance, we employed data augmentation and expansion strategies, enriching the diversity of training data. The model was trained and evaluated on the augmented OASIS1 dataset. Results: The improved Swin-Tiny+ATT model demonstrated significant performance gains across all key metrics on the augmented dataset. Compared to the baseline Swin Transformer, accuracy improved from 84.83% to 87.96%, recall from 89.82% to 91.92%, precision from 85.27% to 91.98%, and the F1 score from 87.26% to 91.89%. These results confirm that the ATT module effectively enhances the model’s ability to capture complex spatial features and identify critical lesion regions. Discussion: The proposed Swin-Tiny+ATT model exhibits strong potential for improving MRI-based classification of Alzheimer’s Disease. The bidirectional spatial attention mechanism successfully directs the model’s focus to relevant anatomical regions, contributing to higher precision and recall. Combined with data augmentation strategies, the approach mitigates class imbalance and enhances generalization. This work provides a promising deep learning framework to support early and accurate diagnosis of AD, with implications for clinical decision-making and personalized treatment planning.
- Open Access
- 23 September 2025
- Published by IMR Press in Reviews in Cardiovascular Medicine
- Vol. 26 (9)
Development and Validation of an Explainable Prediction Model to Assess the Risk of Coronary Artery Disease in Young and Middle-Aged Individuals
Background: There is currently a lack of adequate risk assessment for coronary artery disease in the young and middle-aged population (ages 20–60). This cohort is characterized by limited symptom presentation, low utilization of medical facilities, and challenges in accessing healthcare services. Consequently, these individuals experience difficulties in early disease identification, rendering them susceptible to sudden cardiac death and premature mortality upon the manifestation of symptoms. Data from regular blood and urine tests, as well as questionnaires, are readily available and well-documented across diverse healthcare environments. Hypertension is a notable risk for coronary artery disease within this population. In light of these challenges, we present a risk assessment system for coronary heart disease specifically tailored for young and middle-aged individuals with hypertension, utilizing data derived from blood and urine examinations in conjunction with a brief questionnaire. Methods: The dataset was sourced from the National Health and Nutrition Examination Survey (NHANES) database, covering the years 2005–2019. Following three iterations of feature selection, we identified 26 pertinent features. Subsequently, we developed five predictive models to facilitate large-scale screening for coronary heart disease risk. To enhance the interpretability of our models, we employed SHapley Additive exPlanations (SHAP) to evaluate the individual contributions of each feature. Results: We included 709 patients diagnosed with coronary artery disease and 6409 healthy individuals in our analysis. The results showed that LightGBM exhibited the highest performance (area under the curve (AUC) of 0.93). Conclusions: This study has the potential to facilitate the improved screening of patients with coronary artery disease; we have developed a risk assessment system that is freely accessible to the public: https://prediction-of-coronary-heart-disease-htn-young-adults.streamlit.app/.
- 23 September 2025
- Published by IMR Press in Reviews in Cardiovascular Medicine
- Vol. 26 (9)
Open AccessDevelopment and Validation of an Explainable Prediction Model to Assess the Risk of Coronary Artery Disease in Young and Middle-Aged Individuals
Background: There is currently a lack of adequate risk assessment for coronary artery disease in the young and middle-aged population (ages 20–60). This cohort is characterized by limited symptom presentation, low utilization of medical facilities, and challenges in accessing healthcare services. Consequently, these individuals experience difficulties in early disease identification, rendering them susceptible to sudden cardiac death and premature mortality upon the manifestation of symptoms. Data from regular blood and urine tests, as well as questionnaires, are readily available and well-documented across diverse healthcare environments. Hypertension is a notable risk for coronary artery disease within this population. In light of these challenges, we present a risk assessment system for coronary heart disease specifically tailored for young and middle-aged individuals with hypertension, utilizing data derived from blood and urine examinations in conjunction with a brief questionnaire. Methods: The dataset was sourced from the National Health and Nutrition Examination Survey (NHANES) database, covering the years 2005–2019. Following three iterations of feature selection, we identified 26 pertinent features. Subsequently, we developed five predictive models to facilitate large-scale screening for coronary heart disease risk. To enhance the interpretability of our models, we employed SHapley Additive exPlanations (SHAP) to evaluate the individual contributions of each feature. Results: We included 709 patients diagnosed with coronary artery disease and 6409 healthy individuals in our analysis. The results showed that LightGBM exhibited the highest performance (area under the curve (AUC) of 0.93). Conclusions: This study has the potential to facilitate the improved screening of patients with coronary artery disease; we have developed a risk assessment system that is freely accessible to the public: https://prediction-of-coronary-heart-disease-htn-young-adults.streamlit.app/.
- Open Access
- 24 July 2025
- Published by Wiley in Advanced Intelligent Systems
- Vol. 7 (12)
Enhanced Multitask Learning with Attention Sparse Routing Mechanism for Intelligent Diabetes Prediction
With the rapid advancement in machine learning and artificial intelligence, their application in the medical field has gained increasing prominence, particularly for the early detection and personalized treatment of chronic diseases. This study presents an intelligent approach for early diabetes recognition, based on an encoder–decoder structure and an enhanced multitask learning model. A multilayer self‐attention mechanism is employed to automatically and efficiently extract feature representations, eliminating the limitations of traditional manual feature selection while improving the model's adaptability to complex data. Furthermore, by integrating an enhanced expert module with a sparse routing mechanism, each expert is treated as a weak classifier, leveraging dropout to reduce the network parameters and mitigate overfitting. The sparse routing mechanism dynamically allocates samples to the optimal expert network, achieving a balance between computational efficiency and model accuracy. Experiments are conducted using the TIANCHI precision medicine competition and NHANES data sets. The results demonstrate that this approach significantly improves diabetes risk prediction compared to various baseline models.
- 24 July 2025
- Published by Wiley in Advanced Intelligent Systems
- Vol. 7 (12)
Open AccessEnhanced Multitask Learning with Attention Sparse Routing Mechanism for Intelligent Diabetes Prediction
With the rapid advancement in machine learning and artificial intelligence, their application in the medical field has gained increasing prominence, particularly for the early detection and personalized treatment of chronic diseases. This study presents an intelligent approach for early diabetes recognition, based on an encoder–decoder structure and an enhanced multitask learning model. A multilayer self‐attention mechanism is employed to automatically and efficiently extract feature representations, eliminating the limitations of traditional manual feature selection while improving the model's adaptability to complex data. Furthermore, by integrating an enhanced expert module with a sparse routing mechanism, each expert is treated as a weak classifier, leveraging dropout to reduce the network parameters and mitigate overfitting. The sparse routing mechanism dynamically allocates samples to the optimal expert network, achieving a balance between computational efficiency and model accuracy. Experiments are conducted using the TIANCHI precision medicine competition and NHANES data sets. The results demonstrate that this approach significantly improves diabetes risk prediction compared to various baseline models.
- Open Access
- 14 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 132-141
A Study on Top-level Design and Implementation Pathway of Intelligent Primary Medical and Health Services
- 14 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 132-141
Open AccessA Study on Top-level Design and Implementation Pathway of Intelligent Primary Medical and Health Services
- Open Access
- 14 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 101-113
An Improved Infectious Disease Risk Prediction Model Based on Attention Mechanism
- 14 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 101-113
Open AccessAn Improved Infectious Disease Risk Prediction Model Based on Attention Mechanism
- Open Access
- 11 July 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Multimedia Computing, Communications, and Applications
- Vol. 22 (9) , 1-22
Boosting Semi-Supervised Video Captioning via Learning Candidates Adjusters
Video captioning is a multimodal task on both CV and NLP, whose goal is to automatically obtain the description of video content with natural language statements. Although there are amounts of video data, their annotations with description sentences are very limited. In this paper, we define the semi-supervised video captioning (SSVC) problem in order to improve performance with limited annotations by leveraging the semantic knowledge from both well-annotated samples and no-annotated samples. To address the problem, we introduce a LCA-boosted model (LCABM) for boosting SSVC, where it is first to explore a learnable candidates adjuster to adjust the caption candidates and then treat these adjusted captions as pesudo labels to train the SSVC model with no-annotated samples in reverse. In particular, the model learning is considered as a bi-level optimization problem and solved by an EM-like multi-stage training algorithm. The experiments show the effectiveness of our proposed LCABM, whose performance is comparable and even better than those state-of-the-art fully-supervised methods even with less annotations.
- 11 July 2024
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Multimedia Computing, Communications, and Applications
- Vol. 22 (9) , 1-22
Open AccessBoosting Semi-Supervised Video Captioning via Learning Candidates Adjusters
Video captioning is a multimodal task on both CV and NLP, whose goal is to automatically obtain the description of video content with natural language statements. Although there are amounts of video data, their annotations with description sentences are very limited. In this paper, we define the semi-supervised video captioning (SSVC) problem in order to improve performance with limited annotations by leveraging the semantic knowledge from both well-annotated samples and no-annotated samples. To address the problem, we introduce a LCA-boosted model (LCABM) for boosting SSVC, where it is first to explore a learnable candidates adjuster to adjust the caption candidates and then treat these adjusted captions as pesudo labels to train the SSVC model with no-annotated samples in reverse. In particular, the model learning is considered as a bi-level optimization problem and solved by an EM-like multi-stage training algorithm. The experiments show the effectiveness of our proposed LCABM, whose performance is comparable and even better than those state-of-the-art fully-supervised methods even with less annotations.
- Open Access
- 30 January 2022
- Published by Century Science Publishing Co in Journal of Social Science and Humanities
- Vol. 4 (1)
The Enlightenment of “Social man” Hypothesis Theory to Human Resources Development in Public Sector
- 30 January 2022
- Published by Century Science Publishing Co in Journal of Social Science and Humanities
- Vol. 4 (1)
Open AccessThe Enlightenment of “Social man” Hypothesis Theory to Human Resources Development in Public Sector
- Open Access
- 30 January 2022
- Published by Century Science Publishing Co in Journal of Social Science and Humanities
- Vol. 4 (1)
Analysis of the Influence of Leadership Style on Civil Servant Behavior Performance
- 30 January 2022
- Published by Century Science Publishing Co in Journal of Social Science and Humanities
- Vol. 4 (1)
Open AccessAnalysis of the Influence of Leadership Style on Civil Servant Behavior Performance
- 1 August 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 2266-2270
Design of health decision support system with data analysis and mining functions
By addressing existing problems in the prevailing Chinese health decision support systems (HDSSs) such as the non-uniform system framework design and unclear data analysis functions, the paper designs and summarizes a logical framework and function module configuration for China's HDSSs in conjunction with a detailed discussion of exemplary HDSS data analysis and mining functions in order to provide a reference for the future development of a comprehensive HDSS with data analysis and mining functions.
- 1 August 2015
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 2266-2270
Design of health decision support system with data analysis and mining functions
By addressing existing problems in the prevailing Chinese health decision support systems (HDSSs) such as the non-uniform system framework design and unclear data analysis functions, the paper designs and summarizes a logical framework and function module configuration for China's HDSSs in conjunction with a detailed discussion of exemplary HDSS data analysis and mining functions in order to provide a reference for the future development of a comprehensive HDSS with data analysis and mining functions.