tapas

tapas copied to clipboard

Code example for privacy attacks with completely unknown generator

1

Is there any way to evaluate privacy of synthetic data using your toolbox without any knowledge of the underlying generator? My usecase involves evaluating privacy risk given only the real...

tiadams

fhoussiau

`tapas` currently requires the dataset schema to be defined independently of the dataset. However, in many cases, it might make sense to get the categories from the dataset itself and/or...

fhoussiau

fhoussiau

- This paper proposes holdout metrics for privacy. This is most likely not relevant but worth a look. https://www.frontiersin.org/articles/10.3389/fdata.2021.679939/full - Zanella et al. (better effective epsilon https://arxiv.org/pdf/2206.05199.pdf) - Reconstruction Attacks...

fhoussiau

Add examples *from the paper* in an `examples` folder for easy explanation of how the code works.

2

The goal is to illustrate how to use the various parts of our toolbox together to kickstart code-writing for developers. To begin with, it would be good to have a...

fhoussiau

In some cases, the analyst using `tapas` might not have access to a generator in a usable form. Instead, they have a few synthetic datasets.

fhoussiau

fhoussiau

fhoussiau

I tried to use TAPAS, but I got the following error: ERROR: Package 'tapas' requires a different Python: 3.8.8 not in '=3.9'

FerdoosHN