ProximalAlgorithms.jl

ProximalAlgorithms.jl copied to clipboard

[question] Achieve constraints via return of Inf

3

If you don't want questions in your issues, please just delete. We have complicated constraints that can not (or at least it is not obvious how to) be enforced via...

Maximilian-Stefan-Ernst

Documentation: expand sparse least squares example to other algorithms (Douglas-Rachford)

7

I managed to get most algorithms running on the Sparse regression example, but not the DRLS algorithm. I guess it should be applicable to sparse regression, but what needs to...

patwa67

Remove Zygote dependency by switching to AbstractDifferentiation

11

Hi there! I was wondering if you'd be interested in a PR that abstracts away the Zygote dependency by adopting the formalism from [AbstractDifferentiation](https://github.com/JuliaDiff/AbstractDifferentiation.jl)?

gdalle

Autodiff backends

7

Right now the package falls back to Zygote for computing gradients using automatic differentiation. It would be nice to add support for multiple AD backends (ReverseDiff, Enzyme, Yota, Nabla, Diffractor…...

lostella

Some algorithms code may be improved (and maybe sped up?) by using the 5-arguments version of `mul!` (requires Julia 1.3), see https://docs.julialang.org/en/v1/stdlib/LinearAlgebra/#LinearAlgebra.mul! For example, the code here could probably use...

lostella

Some things that could improve about verbose mode: * Currently, verbose execution of algorithms displays a sequence of unspecified values. Verbose mode should either display a “table header” at the...

lostella

Some of the methods which are usually implemented for Julia iterators are not yet defined on the algorithms iterator types: - `IteratorEltype` - `eltype`

lostella

Some quantities computed in [FBS](https://github.com/kul-forbes/ProximalAlgorithms.jl/blob/master/src/algorithms/forwardbackward.jl#L63) and [PANOC](https://github.com/kul-forbes/ProximalAlgorithms.jl/blob/master/src/algorithms/panoc.jl#L86) could probably be pre-allocated. We should consider doing this in case significant performance improvement is obtained.

lostella

Auto increase gamma

3

Currently, the step size `gamma` parameter can only decrease, eg at https://github.com/JuliaFirstOrder/ProximalAlgorithms.jl/blob/dc0342ca6fa3f6730168b3dcad9a1bd64187d875/src/algorithms/fast_forward_backward.jl#L92 However, sometimes larger step sizes can be performed at some point during iteration. I tried doing `state.gamma *=...

aplavin

Nonmonotone variant of linesearch methods

2

This PR adds the possibility to run algorithms with a linesearch procedure in a nonmonotone variant. In particular, - this feature addresses algorithms such as `PANOC`, `PANOCplus`, `ZeroFPR` and `DRLS`;...

aldma