Learn
Lesson 05 / 14Core20 min

The Optimal Discriminator

Proposition 1: D*(x) = p_data(x) / (p_data(x) + p_g(x))

By the end of this lesson you can

  • Derive the optimal discriminator for a fixed generator
  • Interpret D* as a density ratio
  • See why D* = 1/2 everywhere when p_g = p_data
Visual explainer
Loading lab
Explanation

Before analysing the whole game, the paper asks a simpler question: if the generator stands still, what is the best possible discriminator? The answer has a closed form, and it is the key that unlocks the rest of the theory.

Proposition 1

For G fixed, the optimal discriminator is D*_G(x) = p_data(x) / (p_data(x) + p_g(x))

The analysis in Section 4 is done in a non-parametric setting: D can be any function at all, as if the network had infinite capacity. Keep that assumption in mind.

The Proof, Step by Step

Step 1: write both terms as one integral over x

The second term of V is an expectation over noise z. But G(z) with z ∼ p_z is by definition a sample from p_g, so it can be rewritten as an expectation over x ∼ p_g:

V(G, D) = ∫ p_data(x) log D(x) dx + ∫ p_z(z) log(1 − D(G(z))) dz
         = ∫ [ p_data(x) log D(x) + p_g(x) log(1 − D(x)) ] dx

Step 2: maximise pointwise

D is an arbitrary function, so its value at each x can be chosen independently. To maximise the integral, maximise the integrand separately at every x. At a fixed x, write a = p_data(x), b = p_g(x) and y = D(x). We need to maximise:

f(y) = a · log(y) + b · log(1 − y),    y ∈ [0, 1]

Step 3: one line of calculus

f′(y) = a/y − b/(1 − y) = 0   ⇒   a(1 − y) = b·y   ⇒   y = a / (a + b)

The second derivative, −a/y² − b/(1 − y)², is negative, so this is a maximum. The paper states it as: for any (a, b) ∈ ℝ² \ {(0, 0)}, the function y ↦ a log(y) + b log(1 − y) achieves its maximum in [0, 1] at a/(a+b). Substituting back gives D*_G(x). The paper also notes D need not be defined outside Supp(p_data) ∪ Supp(p_g), since the integrand is zero there.

Reading D* as a Density Ratio

D* compares two densities at the same point. Rearranging:

D*(x) / (1 − D*(x)) = p_data(x) / p_g(x)

So the odds that D* assigns are exactly the ratio of the two densities, and the logit log(D*/(1 − D*)) is the log density ratio. A trained discriminator is a density ratio estimator: without knowing either density, it learns how much more likely x is under the data than under the generator.

  • Where only the data puts mass (p_g(x) = 0): D*(x) = 1.
  • Where only the generator puts mass (p_data(x) = 0): D*(x) = 0.
  • Where the generator over-produces (p_g > p_data): D* < 1/2.
  • Where the generator under-produces (p_g < p_data): D* > 1/2.

This is exactly the signal G needs: D* is high where G should send more samples and low where it should send fewer.

The Perfect Generator Case

If p_g = p_data, then at every x the ratio is p / (p + p) = 1/2. The best possible discriminator is a coin flip. This is the flat blue line at 1/2 in panel (d) of the paper's Figure 1.

Why this matters: once we know D* in closed form, we can substitute it back into V and get a function of G alone. That is the next lesson, and it is where the Jensen-Shannon divergence appears.
Common mistake: assuming a real, finite network actually reaches D*. The proof lets D(x) be set freely at every point. A network with limited capacity, trained for a few steps on finite samples, only approximates it. The theory describes the ideal discriminator, not the one on your GPU.
Interactive code

Edit the code and press Ctrl + Enter to run it in your browser.

pynumpy · matplotlib
Exercises
1

At some point x, p_data(x) = 0.3 and p_g(x) = 0.1. What is D*(x)? What if p_g(x) = 0 there? What should the generator do near this x?

2

A well-trained discriminator outputs D(x) = 0.8 on some input x. Without knowing either density, what can you say about p_data(x) / p_g(x)? Why is this called density ratio estimation?

3

The proof maximises the integrand separately at each x. Which assumption makes this legitimate, and what breaks when D is a small network (for example D(x) = σ(w·x) in the lesson 4 code)?

Key takeaway

For a fixed G, the best discriminator is the ratio p_data / (p_data + p_g); it outputs 1/2 exactly when the generator is perfect.

Up next

Plug D* back into V and a famous divergence appears.