Lagrange multipliers are a go-to tool for anyone who’s worked in optimization.
In this talk we will explore the renowned Karush-Kuhn-Tucker optimality conditions starting from first principles, leveraging visual insights, and gradually progressing to the formal proof in the context of linear constraints.
The talk is based on the blog post: https://mlwithouttears.com/2024/11/06/lagrangian-multipliers-normal-cones-and-kkt-optimality-conditions/