Gretel-synthetics FAQs

What is synthetic data?

How does Gretel synthetics create artificial data?

Is there an architecture diagram?

What kinds of data can I send to Gretel-synthetics?

What are the outputs from Gretel-synthetics?

Can I run gretel-synthetics on premises?

What are gretel-synthetics premium features?

Do I still need to de-identify sensitive data when using gretel-synthetics?

What kinds of privacy protections can Gretel Synthetics help with?

How is Gretel-synthetics differential privacy different from traditional implementations?

How is synthetic data different from the original source data it was trained on?

How many lines of input data do I need to train a synthetic model?

How many columns of training data can I have?

How many epochs should I train my model with?

Does training a synthetic model require a GPU?

What is differential privacy?

How does Gretel-synthetics leverage differential privacy?

How does Gretel-synthetics implement differential privacy?

If my model trained in batches using differential privacy, what is my final epsilon (privacy guarantee)?

What are good epsilon (ε) and delta (δ) values in differential privacy?

How is Stochastic Gradient Descent (SGD) modified to be differentially private?

What does RDP order mean?