Brandon T. Willard
Brandon T. Willard
This addresses #56 in another way; namely, it uses an intermediate/temporary TF graph that mirrors a given meta graph with the meta tensor terms replaced by `Placeholder`s. The temporary TF...
`MetaSymbol.reify` is still pretty stack-driven, which is bad for scaling toward larger graphs, so we should consider implementing a trampolined approach (e.g like `etuples` and `unification`).
An issue arises when reifying meta tensors that are intended to be new to a TF graph, because a tensor with an unspecified name value is currently set—by default—to its...
In some instances, we automatically create logic variables as stand-ins for unspecified meta object components (e.g. #35, [`TFlowMetaOp.outputs`](https://github.com/pymc-devs/symbolic-pymc/blob/master/symbolic_pymc/tensorflow/meta.py#L633)); following #84, we should always produce properly "typed" generic objects and we...
It might be worth investigating a way to partially specify the `NodeDef.attr` `dict`, especially since complete removal/reassignment causes problems for constants (the Numpy constant value is held in that map)....
It would be great if we could automatically transform backend helper/graph constructor functions into meta equivalents. ## The problem These "helper functions" are the standard Python functions found in the...
There's nearly enough in place to start automating the application of proximal methods. I've outlined some of the basic ideas [here](https://brandonwillard.github.io/a-role-for-symbolic-computation-in-the-general-estimation-of-statistical-models.html) and [here](https://brandonwillard.github.io/more-proximal-estimation.html). To recap, this would involve simple identifications...
After the introduction of TensorFlow backend support (i.e. #4 ), we can implement functions to convert to-and-from PyMC4 models.
Let's make some simple examples showing how `symbolic-pymc` can be used to convert a manually constructed model—or components thereof—into its corresponding PyMC3 distribution class. For example, we could show how...
After TensorFlow graph support is added (i.e. #4), we will need to adapt the existing miniKanren relations to TF[P] objects. This might require a core set of linear/tensor-algebraic graph normalization/canonicalization...