Private synthetic data with GANs, autoencoders and minority oversampling

13:50-14:10, January 27 @ 2A

Talk/ Overview

Ioannis will first give an overview of the AI related work carried out at the Data Science Campus of the Office for National Statistics. He will then present the system for synthetic data generation. The goal of this system is to replace real data for the purposes of processing and analysis, which is particularly useful in cases where the real data are sensitive. State-of-the-art methods, such as generative adversarial networks, autoencoders and differential privacy serve as the system's building blocks. Last, Ioannis will introduce some practical applications of this work, such as the generation of synthetic demographic data.

Talk/ Speakers

Ioannis Kaloskampis

Senior Data Scientist, Office for National Statistics UK

Talk/ Slides

Download the slides for this talk.Download ( PDF, 11964.22 MB)

Talk/ Highlights

15:40

Private synthetic data with GANs, autoencoders and minority oversampling

With Ioannis KaloskampisPublished March 12, 2020

AMLD / Global partners