-
Can the API for adding an adjoint rule allow for explicitly specifying the rule for a higher order adjoint?
e.g. `D(sin,x) = v -> v * cos(x)` but I also know that `D(sin,x; n=2) = (v1,v2) -> v2*v1*…
-
2 times the standard correction term from Stratonovich to Ito must be added to the drift function to ensure that we get the proper backwards evolution in the Ito sense.
-
```julia
function loss1(p)
f(x,p,t) = 1
prob = DiscreteProblem(f, 0, (1,10), p)
sol = solve(prob, FunctionMap(scale_by_time = true), saveat=[1,2,3])
return sum(sol)
end
DiffEqFlux.sciml…
-
Calling `getindex` on an adjoint of a sparse array falls back (tested on 1.2 and on master built from source today, both on Linux) to a general `AbstractArray` method, making it unusable for non-scala…
-
Actually `Healpix.jl` does not have support to AD engines like `Zygote.jl` or similar. However, adding support to it is straightforward, as SHT is a linear operator and the mathematical formulas for t…
-
For simple cases we can just differentiate through the DDE solver but if the DDE system contains parameter-dependent C1-discontinuities the forward sensitivities have jump discontinuities which, e.g.,…
-
Hello Halide-Team,
(thanks for putting your efforts in such a neat to tool for more application oriented programmers and scientists!)
I have started to play around with the rather recent featu…
-
Currently, only first order derivatives are provided using Forward Mode Algorithmic Differentiation.
If implemented, the efficiency for first order would be much higher and second order would be poss…
-
From @avik-pal https://github.com/SciML/LinearSolve.jl/pull/449
> Preconditioning. These need to be added to the Adjoint Sensitivity Struct. Can we use a left preconditioner for the forward proble…
-
Why does this fail
```julia
julia> function f3(x)::SMatrix{2,2,Float64,4}
[ x x^2; √x sin(x) ]
end
julia> Zygote.gradient(x->sum(f3(x)),3.3) # ERROR: Need an adjoint for const…