Synthesizing EHR data with GANs

State-of-the-art GAN models for synthesizing patient EHRs with various data types.

(1) C-RNN-GAN

• Paper link: https://arxiv.org/abs/1611.09904
• Code link: https://github.com/olofmogren/c-rnn-gan

(2) RCGAN

• Paper link: https://arxiv.org/abs/1706.02633
• Code link: https://github.com/ratschlab/RGAN

(3) TimeGAN

• Paper link: https://proceedings.neurips.cc/paper/2019/file/c9efe5f26cd17ba6216bbe2a7d26d490-Paper.pdf
• Code link: https://github.com/jsyoon0823/TimeGAN

(4) medGAN

• Paper link: https://arxiv.org/abs/1703.06490
• Code link: https://github.com/mp2893/medgan

(5) synTEG

• Paper link: https://academic.oup.com/jamia/article/28/3/596/6024632
• Code link: https://github.com/allhailjustice/SynTEG

(6) DualAAE

• Paper link: https://academic.oup.com/jamia/article/27/9/1411/5912632
• Code link: https://github.com/donalee/DualAAE-EHR

(7) Privbayes

• Paper link: https://dl.acm.org/doi/10.1145/3134428
• Code link: https://github.com/daanknoors/synthetic_data_generation