18 Jul 2024
Recently, the Department of Science and Technology of Jiangsu Provincial announced the list of funded projects for the 2024 Jiangsu Province Science and Technology Basic Research Program. Dr. Shuihua Wang's exciting project "Multimodal Biomedical Data: Harnessing the Unlabeled Data to Enhance the Biomedical Interpretable Learning" from the School of Science at Xi'an Jiaotong-Liverpool University was competitively selected and approved as a general project.
In summary, this innovative project aims to make a significant and positive impact in biomedical big data. In the field of modern biomedical research, the rapid development of medical imaging omics and high-throughput technology has driven the generation and expansion of biomedical big data. Among them, medical imaging omics, such as traditional X-ray, CT scanning, MRI imaging, as well as more complex imaging techniques such as PET and Ultrasound, utilizes advanced imaging techniques to capture the microscopic features of diseases. Through these technologies, researchers are able to observe and record the internal structure and changes of living organisms with unprecedented accuracy and dimensions, bringing about a massive amount of medical imaging data. Meanwhile, high-throughput technologies such as genome sequencing, proteomics, and metabolomics can automatically analyze thousands of biological samples in a short period of time, revealing a large amount of data about genes, proteins, and other biomolecules. The application of these technologies greatly accelerates the speed of data generation, increases the diversity and complexity of data, and provides clinicians and researchers with abundant resources to explore complex biological processes, disease mechanisms, and potential treatment methods.
However, while the rapid development of these technologies has greatly promoted the development of biomedical big data, it also increases the complexity of data processing and analysis. The main reason for this is that these large-scale biomedical data are characterized by large scale, multiple types, diverse sources, fast generation, huge value but low density, and contain a large amount of unstructured and multi-view data, but a lack of sufficient labeled data. However, machine learning, especially supervised learning algorithms, often rely on a large amount of labeled data to train models, enabling them to accurately identify and predict on new data. However, for real-world applications, many medical imaging data and biomolecular data lack sufficient labels, in other words, there is a lack of known output or classification label, which hinders the widespread application of artificial intelligence (AI) in the biomedical field.
This project aims to use diffusion models and other possible methods to solve the problem of missing modality information in multimodal medical data, while using information decomposition theory to improve the feature extraction process of multimodal data and better develop multimodal machine learning methods to integrate biomedical data from different sources. The aim is to improve the learning ability of artificial intelligence in the biomedical field, not only to understand complex biomedical systems and provide interpretable results, therefore, leading to more accurate and personalized intervention suggestions.
Introduction to Dr. Shuihua Wang
Dr. Shuihua Wang received her PhD degree from Nanjing University in 2017. She was an Assistant Professor at Nanjing Normal University from 2013 to 2018, and a Research Associate at Loughborough University from 2018 to 2019 and a Research Associate and Lecturer at the University of Leicester from 2019 to 2023. She is currently an Associate Professor in the Department of Biological Sciences, School of Science at Xi'an Jiaotong-Liverpool University.
Dr. Wang’s research interests focus on Machine learning, Deep learning, Image processing, Information Fusion, Data Analysis, MRI Sensors, CT Sensors. She is a globally renowned and prolific researcher with numerous high quality published articles in respected peer-reviewed international journals and conferences. During her short time at XJTLU, she has successfully secured several externally funded grants and is in the process of building and supervising a large and strong team of highly qualified Masters and especially PhD students from China and across the world.
Dr. Wang serves as a professional reviewer for many well-reputed international journals and conferences including IEEE Transactions on Neural Networks and Learning Systems, Neurocomputing, Pattern Recognition, Scientific Reports, among others. She serves as Guest Editor of Information Processing and Management, IEEE Journal of Biomedical and Health Informatics, Machine Vision and Applications, IEEE/ACM Transactions on Computational Biology and Bioinformatics and Measurement, Scientific Reports. She serves as Associate Editor of Information fusion, Neural Networks, and IEEE Transactions on Circuits and Systems for Video Technology (TCSVT).
If you are interested to join Dr. Wang’s highly selective and competitive research team, please feel free to contact her at: Shuihua.Wang@xjtlu.edu.cn
Material: Dr Shuihua Wang
Review: Professor John Moraros
18 Jul 2024