ProximalOperators.jl
ProximalOperators.jl copied to clipboard
Proximal operators for nonsmooth optimization in Julia
Hi Lorenzo, I wanted to follow up on this comment: https://github.com/JuliaFirstOrder/ProximalOperators.jl/issues/124#issuecomment-894072191 I think these weighted sums of norms can be useful for Group/Fused Lasso type problems. For the Group Lasso,...
We seem to be missing most of the implementations of `gradient`/`gradient!`. Was this intentional or should I start adding them? Should we define gradients even for functions that are not...
There's a new, robust implementation of the projection onto the exponential cone detailed in [this paper by Henrik Friberg](https://docs.mosek.com/whitepapers/expcone-proj.pdf). This is implemented in MathOptSetDistances.jl. Happy to port over. https://github.com/matbesancon/MathOptSetDistances.jl/blob/1f3e578398ba91fbbcaf805aaf55f9105d858a40/src/projections.jl#L170 Relevant...
Whipped this together earlier. Not sure on the stopping criterion, but it seems reasonable... ```julia using ProximalOperators # Define solvers function NNLS_admm(A, b, x; tol = 1e-9, maxit = 50000,...
May I request for the feature extension to compute the proximal operator of a constrained convex function? For example, a constrained convex function that shows up in low rank factor...
This is a list of (some of the) currently missing functions and calculus rules for proximal mappings, with references. Please comment to add more to the list. **Functions** - [...
As noted several times, code coverage bounces up and down due to randomized tests. One way of avoiding this is to hardcode all cases to be tested: time consuming, but...
Docstrings could be polished by using the `@doc` macro with `raw"..."` strings, see https://juliadocs.github.io/Documenter.jl/stable/man/latex/#Escaping-Characters-in-Docstrings-1. This would get rid of the current abundance of backslashes due to string escaping.
This is just a design idea, but I kinda like it. From the beginning of ProximalOperators, we had `prox!` and `prox` returning both the proximal point `y` *and* the value...
I would like to use the LLR operator, and am considering making a PR. Reference: "Local Low-Rank Matrix Approximation" Joonseok Lee et al. https://static.googleusercontent.com/media/research.google.com/en//pubs/archive/45235.pdf