JuliaTrustworthyAI/CounterfactualExplanations.jl: v0.1.16

Diff since v0.1.15

Closed issues:

• Dealing with count data (#83)
• Choices for default penalty strength (#114)
• Search is slow for high-dimensional data (#142)
• Plotting for multi-dim data is throwing errors (#148)
• Custom print for CounterfactualExplanation looks weird for arrays (#154)
• [Penalty strengths] Investigate ML models for penalty strength tuning (#181)
• [Penalty strengths] Implement penalty strength tuning through ML (#183)
• [Benchmarks] More synthetic datasets (#185)
• benchmark sample ids are messed up using MPI (#271)
• Penalty functions: The way penalty functions like distance(ce, args...; kwargs...) are currently handled makes it difficult to actually supply args... and kwargs... at the generator stage. One way to do this is to use anonymous functions like here but that causes issues with Serialization (#274)
• Penalty functions in composable generators (#275)
• Use grid to more efficiently parallelize benchmarks (#277)
• Add parallelizer for multi-threading (#282)
• The @threads macro is nested which currently leads to issues with certain kwargs... (#285)
• Combine multi-threading and multi-processing (#287)
• Add progress indicators back in (#290)
• Plots (Extension) (#292)
• Deprecate Plots methods (#293)
• distance computations are unnecessarily slow (#296)

Merged pull requests:

• Sorting out #271 (#273) (@pat-alt)
• that'll do for now (#276) (@pat-alt)
• 277 use grid to more efficiently parallelize benchmarks (#279) (@pat-alt)
• Adds support for multi-threading (#283) (@pat-alt)
• Enable support for kw::Symbol in parallelizable functions (#288) (@pat-alt)
• 287 combine multi threading and multi processing (#289) (@pat-alt)
• 290 add progress indicators back in (#291) (@pat-alt)
• 293 deprecate plots methods (#294) (@pat-alt)
• 296 distance computations are unnecessarily slow (#298) (@pat-alt)