Can synthetic tabular data be more accurate?

This new model by researchers at Universidad Politécnica de Madrid combines a Variational Autoencoder with a Gaussian Mixture Model. Results outperform #CTGAN and #TVAE, especially on medical datasets.

👉 synthema.eu/publications/

#SYNTHEMA

CT‑GAN Advances Target Population Synthesis for Urban Planning

CT‑GAN topped single‑variable distribution accuracy; paired with Synthesis Combinatorial Optimization, the hybrid model beat FBS‑CO on joint accuracy. getnews.me/ct-gan-advances-target-p... #ctgan #population #urbanplanning