Towards Scalable Bayesian Learning of Causal DAGs

By Jussi Viinikka, Antti Hyttinen, Johan Pensar, Mikko Koivisto. 2020

46 citations recorded by the source corpus.

We give methods for Bayesian inference of directed acyclic graphs, DAGs, and the induced causal effects from passively observed complete data. Our methods build on a recent Markov chain Monte Carlo scheme for learning Bayesian networks, which enables efficient approximate sampling from the graph posterior, provided that each node is assigned a small number K of candidate parents. We present algorithmic tricks to significantly reduce the space and time requirements of the method, making it feasible to use substantially larger values of K. Furthermore, we investigate the problem of selecting the candidate parents per node so as to maximize the covered posterior mass. Finally, we combine our sampling method with a novel Bayesian approach for estimating causal effects in linear Gaussian DAG models. Numerical experiments demonstrate the performance of our methods in detecting ancestor-descendant relations, and in effect estimation our Bayesian method is shown to outperform existing approaches.

Subfields: Artificial Intelligence, Statistics and Probability, Probabilistic & Bayesian Methods, Causal Inference & Discovery

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Stefano Ermon
2. Yoshua Bengio
3. Simon Lacoste-Julien
4. Aditya Grover
5. Sergey Levine
6. Stefan Bauer
7. Pieter Abbeel
8. Ruslan Salakhutdinov

In Statistics and Probability

1. Stefano Ermon
2. Aditya Grover
3. Chris Cundy
4. Simon Lacoste-Julien
5. Stefan Bauer
6. Yoshua Bengio
7. Tristan Deleu
8. Alex Graves

In Probabilistic & Bayesian Methods

1. Stefano Ermon
2. Yoshua Bengio
3. Aditya Grover
4. Tristan Deleu
5. Simon Lacoste-Julien
6. Stefan Bauer
7. Chris C. Emezue
8. Ant'onio G'ois

In Causal Inference & Discovery

1. Yoshua Bengio
2. Simon Lacoste-Julien
3. Stefan Bauer
4. Aditya Grover
5. Stefano Ermon
6. Tristan Deleu

Citations in the corpus

• Bayesian Structure Learning with Generative Flow Networks (2022, methodology): “We also observe that the maximum a posteriori MEC found by DAG-GFlowNet reaches a higher score than the one found with Gadget, but a lower score than MC 3 ; as a point of reference, the score of the best MEC obtained with GES (Chickering, 2002) is − 10 , 716 .”
• BCD Nets: Scalable Variational Approaches for Bayesian Causal Discovery (2021, methodology): “Furthermore, the ELBO is a natural metric to evaluate convergence, while for GADGET it is unclear how many iterations are required to mix 3 .”