Maths for ML / Notation

H=∇2f\mathbf{H} = \nabla^{2} f

Hessian

CalculusAdvanced

Say it “the Hessian of f”

The matrix of second derivatives — describes the curvature of a function.

Example

∇2(x2+3y2)=[2006]\nabla^2 (x^2+3y^2) = \begin{bmatrix}2&0\\0&6\end{bmatrix}

In code

torch.func.hessian(f)(x)

Where you’ll see it in ML

Second-order optimizers, sharpness of minima, Newton's method.

Don’t confuse it with

Lessons that use it

← All 94 symbols