SMILEMODELING · IMAGING · LEARNING
Home / Research

Research

Our research

We develop machine learning methods for large-scale data in imaging, computer vision, and bioinformatics. Our work combines theoretical analysis with practical algorithms.

01

Algorithms + systems

Scalable learning

Efficient learning and optimization methods designed for data-intensive scientific workloads, with theoretical guarantees and practical performance.

  • Large-scale inverse optimization
  • Structured sparse learning
  • High-performance data processing
02

Images + outcomes

Computational pathology & imaging

Multimodal models that connect whole-slide pathology, medical images, molecular profiles, and clinical outcomes.

  • Whole-slide image analysis
  • Survival prediction
  • MRI reconstruction and segmentation
  • Spatial omics
03

Structure + representation

Graph learning

Deep graph models that learn from relationships in molecular, biomedical, social, and image-omics data.

  • Graph neural networks
  • Geometric representation learning
  • Graph generation
  • Hypergraph learning
04

Molecules + medicine

AI for biology & drug discovery

Machine learning for molecular representation, protein interactions, therapeutic discovery, and biological foundation models.

  • Drug-target interaction
  • Protein and antibody modeling
  • TCR–antigen binding
  • Molecular generation
05

Perception + behavior

Computer vision

Robust visual learning for image and video understanding, multimodal alignment, human behavior, and biomedical applications.

  • Vision-language learning
  • Image and video analysis
  • Domain adaptation
  • 3D/4D modeling

Code and resources

Find our work on GitHub.

We share code and project resources from SMILE projects when they are ready for release.

Visit SMILE on GitHub ↗