Jianming Liang

Dr. Liang's research, funded by an NIH R01 grant and multiple state and local grants, is highly interdisciplinary, meshing several branches of computer science and engineering with medical specialties and social sciences and targeting direct advances in human welfare. In addition to his 100+ peer-reviewed publications, he has been awarded 40 US patents with an additional 50+ patents pending, and his research has led to FDA-approved products.

His research interests include computer-aided diagnosis, medical imaging, computer vision, machine learning, knowledge graphs, causal diagrams, and foundation models. He advocates that foundation models must learn from and be grounded in domain foundations, anchoring their fairness, explainability, and trustworthiness within embedding sociotechnical systems. His lab has been developing medical foundation models that learn from the foundation of medicine, serving the patient's best interest within the healthcare system. 

Publications: Google Scholar

Publicly Released Software:

1. Training foundation models by accruing and reusing knowledge from heterogeneous expert annotations
   https://github.com/JLiangLab/Ark, 2023 (v1), 2024 (v2)
2. Learning foundation models from anatomy in medical imaging via self-supervision
   https://github.com/JLiangLab/Eden, 2023 (v1), 2024 (v2)
3. Large-scale benchmarking and boosting transfer learning for medical image analysis
   https://github.com/JLiangLab/BenchmarkTransferLearning, 2021 (v1), 2024 (v2)
4. Learning anatomically consistent embedding for chest radiography
   https://github.com/JLiangLab/PEAC, 2023
5. Achieving discriminative, restorative, and adversarial learning via stepwise incremental pretraining
   https://github.com/JLiangLab/StepwisePretraining, 2022 (v1) and 2023 (v2)
6. Benchmarking and boosting transformers for medical image classification
   https://github.com/JLiangLab/BenchmarkTransformers, 2022
7. Restoring patch order and appearance for self-supervised medical image analysis
   https://github.com/JLiangLab/POPAR, 2022
8. Developing context-aware instance discrimination for a self-supervised framework
   https://github.com/JLiangLab/CAiD, 2022
9. Uniting discriminative, restorative, and adversarial learning for self-supervised medical image analysis
   https://github.com/JLiangLab/DiRA, 2022 (v1) and 2023 (v2)
10. Evaluating and optimizing deep learning methods for computer-aided diagnosis of pulmonary embolism
    https://github.com/jlianglab/CAD_PE, 2021 (v1) and 2023 (v2)
11. Exploiting the semantics of anatomical patterns for self-supervised learning
    https://github.com/JLiangLab/TransVW, 2021
    https://github.com/JLiangLab/SemanticGenesis, 2020
12. Building a primary source of transfer learning for 3D medical imaging applications
    https://github.com/MrGiovanni/ModelsGenesis, 2019 (v1) and 2020 (v2)
13. Learning fixed points in generative adversarial networks
    https://github.com/JLiangLab/Fixed-Point-GAN, 2019
14. Advancing deep architectures for medical image segmentation
    https://github.com/MrGiovanni/UNetplusplus, 2018 (v1) and 2019 (v2)