Large-Scale Targeted Cause Discovery via Learning from Simulated Data

TL;DR

This paper introduces a scalable machine learning method that efficiently identifies causal factors of a target variable in large systems by learning from simulated data, bypassing full causal graph reconstruction.

Contribution

It presents a novel neural network-based approach that directly infers causal variables with linear complexity, suitable for large-scale systems like gene regulatory networks.

Findings

Outperforms existing methods in large-scale gene network causal discovery

Demonstrates strong generalization across different graph structures and mechanisms

Scales efficiently to thousands of variables

Abstract

We propose a novel machine learning approach for inferring causal variables of a target variable from observations. Our focus is on directly inferring a set of causal factors without requiring full causal graph reconstruction, which is computationally challenging in large-scale systems. The identified causal set consists of all potential regulators of the target variable under experimental settings, enabling efficient regulation through intervention. To achieve this, we train a neural network using supervised learning on simulated data to infer causality. By employing a subsampled-ensemble inference strategy, our approach scales with linear complexity in the number of variables, efficiently scaling up to thousands of variables. Empirical results demonstrate superior performance in identifying causal relationships within large-scale gene regulatory networks, outperforming existing methods that emphasize full-graph discovery. We validate our model's generalization capability across out-of-distribution graph structures and generating mechanisms, including gene regulatory networks of E. coli and the human K562 cell line. Implementation codes are available at https://github.com/snu-mllab/Targeted-Cause-Discovery.