Custom Sensitivities

Legacy Method

Part of the power of Nabla is its extensibility, specifically in the form of defining custom sensitivities for functions. This is accomplished by defining methods for ∇ that specialize on the function for which you'd like to define sensitivities.

Given a function of the form $f(x_1, \ldots, x_n)$, we want to be able to compute $\frac{\partial f}{\partial x_i}$ for all $i$ of interest as efficiently as possible. Defining our own sensitivities $\bar{x}_i$ means that $f$ will be taken as a "unit," and its intermediate operations are not written separately to the tape. For more details on that, refer to the Details section of the documentation.

Intercepting calls

Nabla's approach to RMAD is based on operator overloading. Specifically, for each $x_i$ we wish to differentiate, we need a method for f that accepts a Node in position $i$. There are two primary ways to go about this: @explicit_intercepts and @unionise.

@explicit_intercepts

When f has already been defined, we can extend it to accept Nodes using this macro.

@explicit_intercepts(f::Symbol, type_tuple::Expr, is_node::Expr[, kwargs::Expr])
@explicit_intercepts(f::Symbol, type_tuple::Expr)

As a trivial example, take sin for scalar values (not matrix sine). We extend it for Nodes as

import Base: sin  # ensure sin can be extended without qualification

@explicit_intercepts sin Tuple{Real}

This generates the following code:

begin
    function sin(##367::Node{<:Real})
        #= REPL[7]:1 =#
        Branch(sin, (##367,), getfield(##367, :tape))
    end
end

And so calling sin with a Node argument will produce a Branch that holds information about the call.

For a nontrivial example, take the sum function, which accepts a function argument that gets mapped over the input prior to reduction by addition, as well as a dims keyword argument that permits summing over a subset of the dimensions of the input. We want to differentiate with respect to the input array, but not with respect to the function argument nor the dimension. (Note that Nabla cannot currently differentiate with respect to keyword arguments.) We can extend this for Nodes as

import Base: sum

@explicit_intercepts(
    sum,
    Tuple{Function, AbstractArray{<:Real}},
    [false, true],
    (dims=:,),
)

The signature of the call to @explicit_intercepts here may look a bit complex, so let's break it down. It's saying that we want to intercept calls to sum for methods which accept a Function and an AbstractArray{<:Real}, and that we do not want to differentiate with respect to the function argument (false) but do want to differentiate with respect to the array (true). Furthermore, methods of this form will have the keyword argument dims, which defaults to :, and we'd like to make sure we're able to capture that when we intercept.

This macro generates the following code:

quote
    function sum(##363::Function, ##364::Node{<:Array}; dims=:)
        #= REPL[2]:1 =#
        Branch(sum, (##363, ##364), getfield(##364, :tape); dims=dims)
    end
end

As you can see, it defines a new method for sum which has positional arguments of the given types, with the second extended for Nodes, as well as the given keyword arguments. Notice that we do not accept a Node for the function argument; this is by virtue of using false in that position in the call to @explicit_intercepts.

@unionise

If f has not yet been defined and you know off the bat that you want it to be able to work with Nabla, you can annotate its definition with @unionise.

@unionise code

As a simple example,

@unionise f(x::Matrix, p::Real) = norm(x, p)

For each type constrained argument xi in the method definition's signature, @unionise changes the type constraint from T to Union{T, Node{<:T}}, allowing f to work with Nodes without needing to define separate methods. In this example, the macro expands the definition to

f(x::Union{Matrix, Node{<:Matrix}}, p::Union{Real, Node{<:Real}}) = begin
        #= REPL[9]:1 =#
        norm(x, p)
    end

Defining sensitivities

Now that our function f works with Nodes, we want to define a method for ∇ for each argument xi that we're interested in differentiating. Thus, for each argument position i we care about, we'll define a method of ∇ that looks like:

function Nabla.∇(::typeof(f), ::Type{Arg{i}}, _, y, ȳ, x1, ..., xn)
    # Compute x̄i
end

The method signature contains all of the information it needs to compute the derivative:

• f, the function
• Arg{i}, which specifies which of the xi we're computing the sensitivity of
• _ (placeholder, typically unused)
• y, the result of y = f(x1, ..., xn)
• ȳ, the "incoming" sensitivity propagated to this call
• x1, ..., xn, the inputs to f

A fully worked example is provided in the Details section of the documentation.

Testing sensitivities

In order to ensure correctness for custom sensitivity definitions, we can compare the results against those computed by the method of finite differences. The finite differencing itself is implemented in the Julia package FDM, but Nabla defines and exports functionality that permits checking results against finite differencing.

The primary workhorse function for this is check_errs.

check_errs(
    f,
    ȳ::∇ArrayOrScalar,
    x::T,
    v::T,
    ε_abs::∇Scalar=1e-10,
    ε_rel::∇Scalar=1e-7
)::Bool where T