Ke Yu
[CV]
Contact:
About Me
I am a fifth year Ph.D. student in Intelligent Systems at the University of Pittsburgh, advised by Prof. Kayhan Batmanghelich. My research interests include sample-efficient representation learning (self-, weakly-, semi-supervised learning) and their medical applications.
I received my M.S. degree in statistics from Georgia Institute of Technology. Before that, I completed my undergraduate studies in computer science and electronic engineering at the Huazhong University of Science and Technology. Along with this, I have 7+ years of experience as a data scientist in the healthcare industry.
News
- [2022.06] Paper on boosting the interpretability of clinical risk scores with intervention predictions accepted as a spotlight presentation in DSHealth KDD 2022!
- [2022.05] Paper on weakly-supervised PU learning for anomaly detection accepted at MICCAI 2022!
- [2022.05] I will spend summer 2022 at Google Geo working on active learning for instance segmentation.
- [2022.01] Paper on De novo prediction of cell-drug sensitivies accepted at PSB 2022 (oral).
- [2021.09] Collaborative work with Suvrit's group at MIT about shortcuts in contrastive learning accepted at NeurIPS 2021!
- [2021.03] I will join Google Health as a resech intern this summer.
- [2020.12] Paper on self-supervised representation learning for 3D medical images accepted at AAAI 2021!
- [2020.10] Paper on hierachical drug embedding in hyperbolic space accepted at Journal of Chemical Information and Modeling!
- [2020.10] Three papers [1, 2, 3] accepted at NeurIPS workshops!
- [2019.09] Paper on monitoring ICU mortality risk with LSTM accepted at PSB 2020! (oral)
Education
Publications
* indicates equal contribution
|
Anatomy-guided weakly-supervised abnormality localization in chest x-rays
|
|
Boosting the interpretability of clinical risk scores with intervention predictions
KDD 2022, DSHealth Workshop /
Paper /
|
|
Hierarchical amortized GAN for 3D high resolution medical image synthesis
|
|
De novo prediction of cell-drug sensitivities using deep learning-based graph regularized matrix factorization
PSB 2022 /
Paper /
|
|
Extracting disease-relevant features with adversarial regularization
BIBM 2021 /
Paper /
|
|
Can contrastive learning avoid shortcut solutions?
|
|
Conext Matters: Graph-based Self-supervised Represenation Learning for Medical Images
|
|
Semi-supervised Hierarchical Drug Embedding in Hyperbolic Space
|
|
Context-aware Self-supervised Learning for Medical Images Using Graph Neural Network
NeruIPS 2020, Medical Imaging Workshop /
Paper
|
|
Hyperbolic Molecular Representation Learning for Drug Repositioning
NeruIPS 2020, ML4Molecules Workshop /
Paper
|
|
Hierarchical Amortized Training for Memory-efficient High Resolution 3D GAN
NeruIPS 2020, Medical Imaging Workshop /
Paper
|
|
Monitoring ICU mortality risk with a long short-term memory recurrent neural network
|