Mr. Deng conducts research in biomedical data analysis, particularly in digital pathology and gene expression data. His efforts center on infusing domain knowledge from medical fields into the design of deep learning models to assist in clinical research. While current advancements in deep learning for digital pathology have mainly focused on cancer pathology, Mr. Deng's work also emphasizes non-cancer pathologies, providing a comprehensive understanding of heterogeneous subfields such as renal, gut, and lung pathology, thereby laying a foundation for histopathology research.

Developments in data sharing and computational resources are reshaping digital pathology research. This includes transitioning from small-scale to large-scale analysis, from 2D to 3D imaging, and from single to multi-modality. Classical deep learning methods and expert examination are inadequate for leveraging these trends. Mr. Deng's research principles include building AI-empowered computer vision tools with cross-modal knowledge, advancing clinical research with AI-powered tools, democratizing pathology research, and incorporating foundation models to create a panoramic pathology atlas.