Choose independently and uniformly from . Starting from any , the projected stochastic gradient descent iteration is
Conditional on , , so is an unbiased stochastic gradient. The metric projection onto a closed convex set keeps every iterate in . For a fixed step size, return the Cesàro average .
Projected stochastic gradient descent Created 2026-09-29 Updated 2026-11-05
Projected stochastic gradient descent combines a random gradient step with the Euclidean projection onto a convex set:
In a finite-dimensional Euclidean space, a nonempty closed convex set has a unique projection. This ensures that the iteration is well defined and feasible. For the scalar constraint , the update is simply .
For every feasible , projection is nonexpansive and fixes , so . This inequality underlies convergence estimates after taking conditional expectations. On a general closed nonconvex feasible set, nearest points may not be unique and this convex projection argument does not apply.