I'm a machine learning and applied statistics researcher with a strong interest in building practical, interpretable, and deployable models. My work focuses on solving real-world problems in biomedical, pharmaceutical and other challenging domains.
I’m particularly interested in:
- Uncertainty quantification in neural networks and predictive models
- Federated learning convergency efficiecy and privacy-preserving model development
- Statistical modeling with interpretability emphasizing techniques that are explainable and scientifically grounded
-
🔍 SF-Rx
A multi-output deep neural network framework for predicting drug–drug interactions under realistic, heterogeneous conditions. -
🧠 NeuroRG
An image based deep learning-based pipeline for morphological profiling of CNS cells, facilitating anti-inflammatory compound screening. -
📊 AIDA-attn
An interpretable attention-based deep learning model designed for high-dimensional time-series data, enhancing feature-level understanding. -
🧬 ZIBseq
Implementation of zero-inflated beta regression for metagenomic differential abundance analysis.
This was developed as a methodologically improved extension to an existing tool, and tried to be contributed via pull request to the original project.
I write about statistical models, deep learning approaches, uncertainty estimation, federated learning, and research notes on my blog:
👉 hello-world-jhyu95.tistory.com
- GitHub: @codespermuted
- e-mail: jaehongyu0105@gmail.com