Theodore Papamarkou
Theodore Papamarkou
The following works for gradient!(): ``` using DiffBase, ReverseDiff f(x) = sum(sin, x)+prod(tan, x)*sum(sqrt, x); x = rand(4); result = DiffBase.GradientResult(x); rcfg = ReverseDiff.GradientConfig(x); ReverseDiff.gradient!(result, f, x, rcfg); DiffBase.value(result) DiffBase.gradient(result)...
See JuliaDiff/ForwardDiff.jl#113, cc @jrevels, @mlubin, @eford.
After having completed #155, provide some examples from the [first volume of OpenBUGS](http://www.openbugs.net/Examples/Volumei.html), starting from the [Rats example](http://www.openbugs.net/Examples/Rats.html).
A functional `BasicGibbsJob` has been implemented [here](https://github.com/JuliaStats/Klara.jl/blob/master/src/jobs/BasicGibbsJob.jl). An example of how this Gibbs job can be utilized to sample from a bivariate normal distribution with two correlated random variables is...
Once #155 has been completed, it will be decided whether a higher level DSL will be implemented.
Consider dropping the `Basic` prefix from the various types that carry it, such as `BasicContMuvParameter` for example, in order to reduce the amount of typing. One easy solution is to...
Breaking down #86 to sampler specific issues, to make porting more manageable.
Breaking down #86 to sampler specific issues, to make porting more manageable.
An important part of the code base has been tested already. Make sure that all the code of v0.5.0 is tested thoroughly.
Comments have been added in few places only. Make code more transparent by commenting on it in the source files systematically.