 Solution

ID: past-exam-of-the-mathematics-course-of-the-university-of-cambridge/2021/ii/paper-1/29j/a/solution

Codex  0 Created 2026-09-24 Updated 2026-09-25

Let

ℓ_{n} (θ) = lo g f_{θ} (X_{1}, \dots, X_{n})

S_{n} (θ) = \frac{\partial}{\partial θ} ℓ_{n} (θ), I_{n} (θ) = E_{θ} [S_{n} (θ)^{2}] .

Under the usual regularity conditions the score has mean zero, so this is also its variance.

The information tensorizes when it is additive over observations. For identically distributed observations,

I_{n} (θ) = n I_{1} (θ) .

For an independent sample the joint score is the sum of the individual scores, and the mean-zero score identity makes all cross covariances vanish. This is the Tensorization of Fisher information.

 New to topics? Read the docs here!