New Feature : Classical Data Encoders into TFQ

[karlunho]

A common task for first time users of TFQ is to try and encode classical data into quantum circuits. It would be good if TFQ has some out-of-box encoders based on 1-2 methods found in the literature.

[sassan72]

you mean TFQ doesn't have a method to encode classical data into quantum circuits?

[zaqqwerty]

Embedding classical data into quantum circuits just means that the parameters of the circuit are controlled by the dataset. We have an example of such an embedding using TFQ in our MNIST tutorial, definitely check that out if you're interested in classical data embedding :)

What we don't have is a standard library of techniques from which users can select off-the-shelf solutions, I think that sort of module is what is referred to in this issue.

[lockwo]

I think this would be a good idea, even sort of simple techniques (e.g. basis encoding) might be useful especially for those coming from classical machine learning. There are a number of different techniques that exist in literature but offering some templates like Pennylane, for example would be nice.

[mhucka]

Due to time constraints, we have to prioritize more aggressively. I'm closing this as not essential and unlikely to be pursued. If this is incorrect, please feel free to reopen this issue.