Seth Axen

A number of benchmarks were performed at https://github.com/mlcolab/PathfinderBenchmarks.jl

In principle, if one could specify a different way of choosing the best distribution, then yes, once could maximize ELBO (or some other objective) over some transformation of that distribution...

See https://github.com/invenia/PDMatsExtras.jl/issues/25

Thanks @cscherrer for the analysis and PR! I had considered reusing `u` before but ultimately decided not to in the interest of allowing the user to inspect all intermediates (they...

Keeping in mind #15, the goal is ultimately to allow the user to provide an object that configures an optimization procedure over the trace of distributions, with the default being...

I have mixed feelings on this. We do have a resample and restart mechanism, so we could suppress errors. But the source of the error matters. We should detect and...

I just tried to test this with the latest versions of all packages on Julia v1.10, but it failed to load because ModelingToolkit failed to precompile. Do you happen to...

Here's an MWE for this error: ```julia using LogDensityProblems, Pathfinder, Transducers struct StdNormalProblem n::Int end LogDensityProblems.capabilities(::Type{StdNormalProblem}) = LogDensityProblems.LogDensityOrder{1}() LogDensityProblems.dimension(prob::StdNormalProblem) = prob.n LogDensityProblems.logdensity(prob::StdNormalProblem, x) = -sum(abs2, x) / 2 LogDensityProblems.logdensity_and_gradient(prob::StdNormalProblem, x)...

This new failure in Transducers might be related: https://github.com/JuliaFolds/Transducers.jl/issues/557

The new issue to follow is https://github.com/JuliaFolds2/Transducers.jl/issues/10