A Lightweight Gradient-based Causal Discovery Framework with Applications to Complex Industrial Processes

TL;DR

This paper introduces GRNGC, a flexible, gradient-regularized neural causality framework that efficiently infers causal relationships in complex systems, outperforming existing methods in simulations and real-world datasets.

Contribution

The paper presents a novel gradient regularization approach for neural Granger causality that reduces computational costs and enhances flexibility across different neural network architectures.

Findings

GRNGC outperforms existing baselines in simulations.

It significantly reduces computational overhead.

Effective in real-world gene regulatory network reconstruction.

Abstract

With the advancement of deep learning technologies, various neural network-based Granger causality models have been proposed. Although these models have demonstrated notable improvements, several limitations remain. Most existing approaches adopt the component-wise architecture, necessitating the construction of a separate model for each time series, which results in substantial computational costs. In addition, imposing the sparsity-inducing penalty on the first-layer weights of the neural network to extract causal relationships weakens the model's ability to capture complex interactions. To address these limitations, we propose Gradient Regularization-based Neural Granger Causality (GRNGC), which requires only one time series prediction model and applies $L_{1}$ regularization to the gradient between model's input and output to infer Granger causality. Moreover, GRNGC is not tied to a specific time series forecasting model and can be implemented with diverse architectures such as KAN, MLP, and LSTM, offering enhanced flexibility. Numerical simulations on DREAM, Lorenz-96, fMRI BOLD, and CausalTime show that GRNGC outperforms existing baselines and significantly reduces computational overhead. Meanwhile, experiments on real-world DNA, Yeast, HeLa, and bladder urothelial carcinoma datasets further validate the model's effectiveness in reconstructing gene regulatory networks.