Will Tebbutt

Ah, so you need to add an adjoint for `literal_getproperty`, rather than either `getproperty` or `getfield`. If `Taylor1` doesn't already implement a method of `getproperty`, you'll (presumably) need to implement...

Ah yeah, I reckon that dispatch on the field is a Zygote-specific thing -- you're correct that `getproperty` just has access to the symbol (as far as I'm aware). See...

I guess the culprit is the discrepancy between the version of the rule DiffRules and ChainRules - https://github.com/JuliaDiff/DiffRules.jl/blob/c90e0bd88e908dc414bf11bb21fc4130800721b0/src/rules.jl#L88 - https://github.com/JuliaDiff/ChainRules.jl/blob/4ad975826ea0639ad709aeb36cc5051b6bf82eaa/src/rulesets/Base/fastmath_able.jl#L170 In the latter we adopt a different convention from the...

Unclear to me -- the "subgradient" convention adopted in ChainRules works perfectly in the KernelFunctions use-case. Also, it's what we do for lots of other things (ReLUs, abs, etc). My...

> Hmm, using DiffRules seems to be more "correct". Thoughts? Hmm, I interpreted this as meaning that you believe the convention adopted in DiffRules is more appropriate than the one...

Thanks for opening this @parikshit-pareek . Could you please provide a full copy of your stack trace so that I can get a better idea of where the problem might...

Ahhh It's an AD issue. The way to solve it is to implement `_map_meanfunction(::CustomMean{typeof(g)}, ::ColVecs)` using something like the matrix operations that you've described above. e.g. ```julia function AbstractGPs._map_meanfunction(::CustomMean{typeof(g)}, x::ColVecs)...

@mjp98 if you could confirm that this _seems_ to fix things on your end, I'll merge

I think it's fair to say that we've avoided taking this implementation task seriously until now because we've not had a complelling use-case -- about the most that we've been...

Ahh I see. This definitely does indeed sound like an interesting use-case. Probably the best way forward is to first consider how equality tests would be included in our standard...