Symbolic Differentiation

Symbolic Differentiation#

Symbolic differentiation mimics programmatically what you do by hand when calculating derivatives. It works by building up a representation of the mathematical expression that you want to differentiate. This is called an expression tree.
Expression trees are made out of nodes that can be either operators or operands. Operators are things like +, -, *, /, or functions like sin, cos, exp, log, etc. Operands are things like numbers, variables, or other expressions.

For example, the expression:

\[ f(x) = 1 + x^2 \]

would be represented by the following expression tree. We have highlighted with operators with red and operands with blue.

Hide code cell source
import graphviz

g = graphviz.Digraph(edge_attr={'arrowhead': 'none'})

g.node('+', color='red')
g.node('sqr', color='red')
g.node('1', color='blue')
g.node('x', color='blue')
g.edge('+', '1')
g.edge('+', 'sqr')
g.edge('sqr', 'x')
g
../_images/7914407e0e3b0494a87738d11e07951a58656b34db2205d020f33154cce4bf24.svg

Here is an expression tree for:

\[ f(x,y) = \left(\frac{\sin x}{x}\right)^2 + y^3. \]
Hide code cell source
g = graphviz.Digraph(edge_attr={'arrowhead': 'none'})

g.node('+', color='red')
g.node('sqr', color='red')
g.node('sin', color='red')
g.node('pow', color='red')
g.node('/', color='red')
g.node('3', color='blue')
# make two different nodes for x
g.node('x1', color='blue', label='x')
g.node('x2', color='blue', label='x')
g.node('y', color='blue')

g.edge('sin', 'x1')
g.edge('/', 'sin')
g.edge('/', 'x2')
g.edge('sqr', '/')
g.edge('+', 'sqr')
g.edge('+', 'pow')
g.edge('pow', '3')
g.edge('pow', 'y')

g
../_images/86df027f176611d1b97a43b45c631485b0037ba264fc832d4c3bc0d23900e67b.svg

You get the idea.

One of the first thing that we would like to do is evaluate the expression. Here is pseudocode for how to evaluate an expression tree:

    def evaluate(node):
        if node is an operator:
            evaluate the operands
            apply the operator to the operands
            return the result
        else:
            return the value of the operand

Similarly, we can write generic code to differentiate an expression tree with respect to a variable. To write such code, we need to know the derivative of each operator and the chain rule. For example, the derivative of the addition operator is:

\[ \partial_x (f + g) = \partial_x f + \partial_x g. \]

Here we are using the notation \(\partial_x f\) to mean the gradient of \(f\) with respect to the variable \(x\), i.e., \(\partial_x f = \frac{\partial f}{\partial x}\).

Similarly, for the multiplication operator:

\[ \partial_x (f g) = f \partial_x g + g \partial_x f. \]

And so on for all operators.

Equipped with these basic operations, we can write a generic function to differentiate an expression tree with respect to a variable:

    def differentiate(node, variable):
        if node is an operator:
            differentiate the operands
            apply the derivative rule for the operator to the operands
            return the result
        else:
            if node is the variable:
                return 1
            else:
                return 0

If you want to learn more about this, I suggest you read Structure and Interpretation of Computer Programs (SICP) by Abelson and Sussman.

The Python package sympy is a library for symbolic mathematics. Let’s use it to build these trees.

import sympy
from sympy.abc import x, y

f = (sympy.sin(x) / x) ** 2 + y ** 3

f
\[\displaystyle y^{3} + \frac{\sin^{2}{\left(x \right)}}{x^{2}}\]

Here is the expression tree as it is internally represented by sympy.

sympy.srepr(f)
"Add(Pow(Symbol('y'), Integer(3)), Mul(Pow(Symbol('x'), Integer(-2)), Pow(sin(Symbol('x')), Integer(2))))"

You can evaluate the expression tree:

f.evalf(subs={x: 1, y: 2})
\[\displaystyle 8.70807341827357\]

Or you can make a python function that evaluates the expression tree at any point:

f_func = sympy.lambdify([x, y], f)

f_func(1, 2)
8.708073418273571

And of course, you can differentiate the expression tree:

# Differentiate with respect to x
pxf = f.diff(x)
pxf
\[\displaystyle \frac{2 \sin{\left(x \right)} \cos{\left(x \right)}}{x^{2}} - \frac{2 \sin^{2}{\left(x \right)}}{x^{3}}\]

And you can evaluate that anywher you want:

pxf_func = sympy.lambdify([x, y], pxf)

pxf_func(1, 2)
-0.5068494097214605

There are a few problems with this approach. First, it is hard to differentiate through loops and conditionals symbolically. This restricts quite a bit the kind of functions that you can differentiate. Second, the expression trees can get quite large and complicated very quickly. This is called the expression swell problem.

Here is an example of expression swell. We use the soft relu function nested a few times:

w, b = sympy.symbols('w b')
# Soft relu
f = sympy.log(1 + sympy.exp(w * x + b))
f
\[\displaystyle \log{\left(e^{b + w x} + 1 \right)}\]

Here is the gradient of this function with respect to \(w\):

f.diff(w)
\[\displaystyle \frac{x e^{b + w x}}{e^{b + w x} + 1}\]

Now let’s nest the soft relu function a few more times:

f2 = sympy.log(1 + sympy.exp(w * f + b))
f2
\[\displaystyle \log{\left(e^{b + w \log{\left(e^{b + w x} + 1 \right)}} + 1 \right)}\]

Here is the derivative now:

f2.diff(w)
\[\displaystyle \frac{\left(\frac{w x e^{b + w x}}{e^{b + w x} + 1} + \log{\left(e^{b + w x} + 1 \right)}\right) e^{b + w \log{\left(e^{b + w x} + 1 \right)}}}{e^{b + w \log{\left(e^{b + w x} + 1 \right)}} + 1}\]

Okay, one more time to make the point:

f3 = sympy.log(1 + sympy.exp(w * f2 + b))
f3
\[\displaystyle \log{\left(e^{b + w \log{\left(e^{b + w \log{\left(e^{b + w x} + 1 \right)}} + 1 \right)}} + 1 \right)}\]

Differentiate:

f3.diff(w)
\[\displaystyle \frac{\left(\frac{w \left(\frac{w x e^{b + w x}}{e^{b + w x} + 1} + \log{\left(e^{b + w x} + 1 \right)}\right) e^{b + w \log{\left(e^{b + w x} + 1 \right)}}}{e^{b + w \log{\left(e^{b + w x} + 1 \right)}} + 1} + \log{\left(e^{b + w \log{\left(e^{b + w x} + 1 \right)}} + 1 \right)}\right) e^{b + w \log{\left(e^{b + w \log{\left(e^{b + w x} + 1 \right)}} + 1 \right)}}}{e^{b + w \log{\left(e^{b + w \log{\left(e^{b + w x} + 1 \right)}} + 1 \right)}} + 1}\]

You get it now.