ManySpace
Türkçe Sign in Sign up

AI & MLRecommendedAIML-PGM-010

Variational Inference and the ELBO

Turning inference into optimisation: fit a simple distribution by maximising a lower bound on the evidence.

Difficulty
3 / 5
Importance
4 / 5
Study time
1–2 h
Kind of work
Mixed

Why it matters

The learner can decompose log evidence into ELBO plus KL and explain why reverse-KL fitting underestimates posterior spread.

The route to this topic

Route AIML-PGM-010
Natural Numbers What Is a Model? Variables & Constants Features and Targets Solving Linear Equations Function Notation & Composition Linear Regression I: Prediction and Residuals k-Means Complex Numbers Loss and Fitting: Squared Error Basic Ratios Unit Circle Objective Functions and Empirical Risk Polynomial Functions Likelihood Intuitive Limits Continuity Definition of the Derivative Differentiation Rules Maximum Likelihood Estimation Bayesian Reasoning Chain Rule Antiderivatives & Indefinite Integrals Mixture Models Intuition Derivatives of Transcendental Functions Riemann Sums & Definite Integral Expectation-Maximization Fundamental Theorem of Calculus Integration by Parts Sample Spaces & Events Random Variables & Distributions Entropy & Mutual Information Variational Inference and the ELBO