strukturalarning
Strukturalarning is the process of inferring the structure of a model or representation from data. It focuses on discovering how variables relate to each other, beyond estimating fixed parameters. The term appears in multiple languages as a variant of "structure learning" and is used across statistics, machine learning, and artificial intelligence.
In many contexts, strukturalarning aims to recover a graph that encodes conditional independencies or causal relationships
Methods are typically categorized into constraint-based approaches, score-based approaches, and hybrids. Constraint-based methods test conditional independencies
A related subfield is neural architecture search, where the structure of a neural network is learned or
Applications span genomics, systems biology, finance, social science, and beyond, including automated model selection and causal