By leveraging the principles of systems biology, graph theory, and gene expression, we employ feature selection techniques to seamlessly integrate clinical data, enabling us to achieve highly accurate prognostic predictions.

In medicine and biology, datasets often face challenges of insufficient sample sizes and large missing values. To address this issue, we have developed and applied generative models, semi-supervised and self-supervised learning techniques, etc. Additionally, we have introduced Bayesian neural networks to enhance the stability and robustness of our models.