-
-
There is almost everything needed in pytorch to write a probabilistic programming (PP) library (see e.g. [Pyro](http://pyro.ai)) and it looks like a good exercise. Plus, if it works well, it could for…
-
What would it take to build out a SM.jl representation of a PPL like http://www.zenna.org/Omega.jl/latest/model/ or https://github.com/cscherrer/Soss.jl?
-
## 一言でいうと
確率モデルを記述するためのフレームワークEdwardの論文。確率変数をつないでグラフ構造を構築するように記載でき、TensorFlowのグラフと統合し爆速で動作可能。VAEやGANなど、確率モデルと統合されたようなモデルが書きやすくなる。
### 論文リンク
https://arxiv.org/abs/1701.03757
### 著者/所属機関
…
-
https://arxiv.org/abs/1701.03757
TMats updated
6 years ago
-
I'm going to log my activities regarding probabilistic programming languages on this issue page.
This will lead to a new BNN model using a more powerful Python library, i.e. **Pyro**.
I'll also put…
-
I could probably do a session on this, but so could Jonny Law
-
Have you considered probabilistic programming for error propagation? Automatic differentiation (together with the delta method from statistics) is pretty nice, too -- and popular courtesy of backpropa…
-
# Probabilistic Programming Languages for Large Scale LFI
Frameworks like [Pyro](https://github.com/pyro-ppl/pyro) and [TensorFlow Probability](https://www.tensorflow.org/probability) are powerful …
EiffL updated
4 years ago
-
In ordinary probabilistic models with a fixed set of variables, this is easy to understand. But when you have different sets of variables in different branches of the program, it can be hard to unders…