I-FLOP: Fast Learning of Order and Parents from Interventional Data
Machine Learning
Summary
The authors extended an existing algorithm called FLOP that learns cause-and-effect relationships from observational data, so it can now also work with interventional data where some variables are deliberately changed. They adapted a scoring method to fit FLOP’s fast updating process. Their new version, I-FLOP, can accurately find the underlying causal structure as data grows large. They tested I-FLOP against other similar algorithms and found it often works better and faster on both real and simulated intervention data.
FLOP algorithmcausal structure learninginterventional dataBIC scoreMarkov equivalence classdirected acyclic graph (DAG)Cholesky decompositioncausal inferenceiterative score updates
Authors
Liuting Chen, Alex Markham
Abstract
We extend the FLOP (fast learning of order and parents) algorithm recently proposed by Wienöbst et al. (2026) from observational to interventional data. In particular, we use the interventional BIC score of Hauser and Bühlmann (2012), adapting it to be used with the iterative Cholesky-based score updates that are partly responsible for FLOP's speed. We show that, in the sample limit, I-FLOP recovers a DAG in the same interventional Markov equivalence class as the data-generating DAG. We compare I-FLOP to existing causal structure learning algorithms on real and simulated interventional data, where it performs favorably in terms of both performance and run time.