Consistent DAG selection for Bayesian Causal Discovery under general error distributions

Anamitra Chaudhuri, Anirban Bhattacharya, Yang Ni · 2025

Read Consistent DAG selection for Bayesian Causal Discovery under general error distributions by Anamitra Chaudhuri, Anirban Bhattacharya, Yang Ni (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-00993.txt.

Description

Anamitra Chaudhuri, Anirban Bhattacharya, and Yang Ni study causal structure learning in a 2025 arXiv paper. The goal is to recover the true directed graph of causes among variables from observational data. Their setting allows errors that are not Gaussian, which widens what can be identified.

They propose a Bayesian hierarchical model over linear structural equations. Using the shape of the errors, they characterize the best class of graphs that data can pin down. Under mild assumptions, the posterior weight on that class grows to certainty as samples grow. Simulations and examples back the consistency claim.

Catalog record

CREATORAnamitra Chaudhuri, Anirban Bhattacharya, Yang Ni
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE200 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-00993

Browse the full library · About the Tower of Babel