Tabular Data Synthesis Using Generative Models
The project aims on supporting research community by reproducing papers without open-source code.
Project GitHub: Not Ready.
The project is conducted by my 2nd year student, Abramov Georgiy, at HSE (Coursework) in 24/25 academic year.
Modern machine and deep learning systems require a lot of data to train on. However, the datasets might be unbalanced or have missing/private information. This may lead to biased computational models that share sensitive data without permission or have poor performance. One way of overcoming this issue is to synthesize data. In this project, the goal is to explore the existing solutions for generating tabular data and reproduce some of them.