Sebastian Banert

The algorithms of Boţ/Hendrich (primal-dual Douglas–Rachford), Condat, Vũ, and Chambolle/Pock are very similar to each other and can all be called primal-dual forward-backward algorithms. They differ by their choice of...

In this jargon, “forward” is for the gradient step(s), and “backward” is for the proximal step(s). This is derived from forward and backward Euler, since forward and backward steps result...

I checked the code, and it yields exactly the vanilla forward-backward algorithm if the lists `g`, `L`, and `sigma` are empty. Wouldn’t it be enough to mention this behaviour in...

Would it be feasible to use Tensorflow (or another learning framework) as a backend to build the computation graph and executing? Another useful aspect of this would be that you...

What I'm concerned about is that not having an abstract `Element` class makes extending ODL harder when it comes to other discretisations which do not allow the representation as a...

I can't reproduce this, either. My output: ``` 2019-01-29 20:38:15.153319: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA...

```conda install -c anaconda tensorflow``` gives the result ```Collecting package metadata: done Solving environment: done # All requested packages already installed. ``` I still can't reproduce the error, working with...

It is rather obvious once you look at them, because the decomposition decouples the coordinates, and each coordinate appears only in at most one absolute value. For the function f(x,...

I don’t know what applies to the English Wikipedia, but in the German one the article would be deleted almost instantly because it violates the relevance criteria (for the English...

The derivative of a linear operator is the operator itself, so this surely doesn't harm anything. On the other hand, the code in this form also works for a non-linear...