Deep Convolutional Neural Networks for Pairwise Causality

TL;DR

This paper introduces a novel approach using deep convolutional neural networks to infer pairwise causal relations from scatter plot images, outperforming previous methods especially with limited data, and enhancing causal discovery capabilities.

Contribution

The paper applies CNNs to causal inference by transforming attribute pairs into images, demonstrating improved performance over gradient-boosted classifiers, especially with less training data.

Findings

CNN approach outperforms GBC in low-data scenarios

Ensemble of CNN and GBC yields significant accuracy improvements

Pre-trained CNNs can be broadly applied to causal discovery tasks

Abstract

Discovering causal models from observational and interventional data is an important first step preceding what-if analysis or counterfactual reasoning. As has been shown before, the direction of pairwise causal relations can, under certain conditions, be inferred from observational data via standard gradient-boosted classifiers (GBC) using carefully engineered statistical features. In this paper we apply deep convolutional neural networks (CNNs) to this problem by plotting attribute pairs as 2-D scatter plots that are fed to the CNN as images. We evaluate our approach on the 'Cause- Effect Pairs' NIPS 2013 Data Challenge. We observe that a weighted ensemble of CNN with the earlier GBC approach yields significant improvement. Further, we observe that when less training data is available, our approach performs better than the GBC based approach suggesting that CNN models pre-trained to determine the direction of pairwise causal direction could have wider applicability in causal discovery and enabling what-if or counterfactual analysis.