Contextual Directed Acyclic Graphs

arXiv:2310.15627v2 Announce Type: replace-cross
Abstract: Estimating the structure of directed acyclic graphs (DAGs) from observational data remains a significant challenge in machine learning. Most research in this area concentrates on learning a single DAG for the entire population. This paper considers an alternative setting where the graph structure varies across individuals based on available "contextual" features. We tackle this contextual DAG problem via a neural network that maps the contextual features to a DAG, represented as a weighted adjacency matrix. …

abstract arxiv challenge cs.lg dag data features graph graphs machine machine learning paper population research stat.ml type