Alexander Viand

@asraa I tried addressing your nit, but unfortunately it seems like it's a limitation of how the 1toN conversion framework works. Modulo "feedback/copybara — Google internal checks FAILED" this should...

@asraa I reabsed this onto main and moved the tests to `tests/transforms/convert_tensor_to_scalars` in keeping with the recent test reorg. EDIT: looks like I missed a `lib/Conversion` somewhere... EDIT2: should be...

> Wouldn't it also be beneficial to vectorize plaintext values when we're performing things like dot products with plaintext inputs? Or this ends up being bucketed to the smarter pre-heir-simd-vectorizer...

> I'd like to talk at the HEIR WG meeting this week on the LWE type - we know we need to revamp the encoding attribute to accommodate prime modulus...

Thanks for the great writeup! It seems like this isn't "real" open source and the most interesting bits are indeed missing. They have a reference to a "[Docker](https://hub.docker.com/repository/docker/zhaozian/heir/general) version of...

See #1016 for a continuation of this PR

Let me also add a list of information I think we need from a BGV/BFV ciphertext (assuming we follow Kim et al. and do a unified implementation) for the bgv->lwe/poly...

I've played around a bit with this now, and come up with the following design proposal: The `lwe.rlwe_ciphertext` no longer has a single `lwe.rlwe_params` attribute, but instead a series of...

> @AlexanderViand-Intel I suspect there is a particular part of this encoding business that is the most uncertain. Can you decide on which and we can iron that out first?...

My recommendation would be to take a look at [.devcontainer/cmake/setup_mlir.sh](https://github.com/google/heir/blob/main/.devcontainer/cmake/setup_mlir.sh) and [.devcontainer/cmake/setup_heir.sh](https://github.com/google/heir/blob/main/.devcontainer/cmake/setup_heir.sh) which (modulo paths being a bit different) should hopefully reflect "best practice". @ai-mannamalai let me know if they...