NICEFA Logo
nicefa.
Library
←Module

Jensen's Inequality for Convex Functions

Exploring the cinematic intuition of Jensen's Inequality for Convex Functions.

Visualizing...

Our institutional research engineers are currently mapping the formal proof for Jensen's Inequality for Convex Functions.

Apply for Institutional Early Access →

The Formal Theorem

Let I I be an interval in R \mathbb{R} and let f:I→R f: I \to \mathbb{R} be a convex function. If x1,x2,…,xn x_1, x_2, \dots, x_n are points in I I , and λ1,λ2,…,λn \lambda_1, \lambda_2, \dots, \lambda_n are non-negative weights such that ∑i=1nλi=1 \sum_{i=1}^n \lambda_i = 1 , then:
f(∑i=1nλixi)≤∑i=1nλif(xi) f\left( \sum_{i=1}^n \lambda_i x_i \right) \leq \sum_{i=1}^n \lambda_i f(x_i)
For a random variable X X taking values in I I , the inequality takes the form:
f(E[X])≤E[f(X)] f(\mathbb{E}[X]) \leq \mathbb{E}[f(X)]

Analytical Intuition.

Imagine a bowl-shaped surface representing a convex function f f . If you place several weights λi \lambda_i at various points xi x_i along the floor of this bowl, the center of mass of those weights—the expected value—necessarily lands at a point whose vertical position on the curve is lower than the weighted average of the individual heights. Visually, the chord connecting any two points on a convex curve lies entirely above the curve. Jensen’s Inequality generalizes this geometric observation to multiple points: the 'average of the function's outputs' will always be greater than or equal to the 'function of the average of the inputs'. Think of it as a manifestation of the 'penalty' for spreading out inputs over a curved landscape; because the slope of the curve is non-decreasing, extreme fluctuations in X X yield disproportionately higher outputs for f(X) f(X) , pulling the average E[f(X)] \mathbb{E}[f(X)] above the value the function would take at the mean input, f(E[X]) f(\mathbb{E}[X]) .
CAUTION

Institutional Warning.

Students often swap the inequality direction. A helpful mnemonic is that for convex (bowl-shaped) functions, the average of the outputs is always 'heavier' or 'higher' than the output of the average input, hence f(mean)≤mean(f) f(\text{mean}) \leq \text{mean}(f) .

Academic Inquiries.

01

What happens if the function is strictly convex?

If f f is strictly convex, equality holds if and only if all xi x_i are identical (or the random variable X X is a constant).

02

Does Jensen's Inequality apply to concave functions?

Yes, but the inequality sign flips. For a concave function, f(E[X])≥E[f(X)] f(\mathbb{E}[X]) \geq \mathbb{E}[f(X)] .

Standardized References.

  • Definitive Institutional SourceRudin, W., Real and Complex Analysis

Related Proofs Cluster.

Intermediate

Weierstrass Extreme Value Theorem: Guaranteeing Existence of Optima

Exploring the cinematic intuition of Weierstrass Extreme Value Theorem: Guaranteeing Existence of Optima.

Intermediate

Local Optima are Global Optima for Convex Functions

Exploring the cinematic intuition of Local Optima are Global Optima for Convex Functions.

Intermediate

Hessian Matrix and Second-Order Optimality Conditions

Exploring the cinematic intuition of Hessian Matrix and Second-Order Optimality Conditions.

Intermediate

Proof that the Intersection of Convex Sets is Convex

Exploring the cinematic intuition of Proof that the Intersection of Convex Sets is Convex.

Institutional Citation

Reference this proof in your academic research or publications.

NICEFA Visual Mathematics. (2026). Jensen's Inequality for Convex Functions: Visual Proof & Intuition. Retrieved from https://nicefa.org/library/fundamentals-of-optimization/jensen-s-inequality-for-convex-functions

Dominate the Logic.

"Abstract theory is just a movement we haven't seen yet."

Subscribe for Full ProofsEarly Access