Higher order forward mode autodiff

JuliaStats / Klara.jl

MCMC inference in Julia

Other

167 stars 38 forks source link

Higher order forward mode autodiff #112

Closed papamarkou closed 7 years ago

papamarkou commented 8 years ago

Once the samplers that require higher order derivatives (such as SMMALA, MMALA, RMHMC) have been ported, the autodiff-related parameter constructors need to be updated to allow calculating the required higher order derivatives via forward mode autodiff.

papamarkou commented 7 years ago

Sorted from v0.8.0 onwards, at least for SMMALA. On the other hand, MMALA and RMHMC require third order derivatives, which are not supported by ForwardDiff.