Adding Training data which distribution differs from Dev/Test sets

[manifest]

@saiman

Thanks. would you mind please explain it more.

We build neural networks to model the real-world data distribution. If we already know that distribution, we don’t need neural networks. In other words, if we knew everything that may happen, there would be nothing to predict

@andreygizatullin

to learn a distribution of the dataset you might use variational autoencoders

Evaluation of generative models quite a problem itself