Homepage › Solution manuals › Kevin P. Murphy › Machine Learning: a Probabilistic Perspective › Exercise 13.13 - Lower bounds to convex functions
Exercise 13.13 - Lower bounds to convex functions
Answers
Since is a convex function, then any hyperplane that is tangent to an arbitrary point has the entire above it. Denote the normal of the hyperplane at by , then the hyperplane intersect with at with:
And we have:
This assertation holds for arbitrary , hence:
For more details, refer to Rigorous Affine Lower Bound Functions for Multivariate Polynomials and Their Use in Global Optimisation.