Shylock: Causal Discovery in Multivariate Time Series based on Hybrid Constraints

2025-10-24

AI Summary

Existing multivariate time series (MTS) causal discovery methods rely on strong assumptions, hand-crafted priors, or large amounts of labeled data, leading to overfitting and poor generalization under data-scarce conditions. This paper proposes Shylock: an end-to-end causal discovery framework applicable to both few-shot and standard-data regimes. Its core innovation is a global-local hybrid constraint mechanism, integrating grouped dilated convolutions with shared-kernel design—dramatically reducing parameter count (exponentially) while enhancing lag-aware temporal modeling and cross-variable information sharing. Coupled with a novel differentiable MTS generator and differentiable causal graph learning, Shylock improves robustness in low-data settings. Extensive experiments on multiple public benchmarks and synthetic datasets demonstrate that Shylock significantly outperforms state-of-the-art methods, especially in few-shot scenarios, achieving marked gains in causal identification accuracy. The code is publicly released as the Tcausal library and integrated into the EarthDataMiner platform.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Security and Privacy: Data transparency and provenance

Abstract

Causal relationship discovery has been drawing increasing attention due to its prevalent application. Existing methods rely on human experience, statistical methods, or graphical criteria methods which are error-prone, stuck at the idealized assumption, and rely on a huge amount of data. And there is also a serious data gap in accessing Multivariate time series(MTS) in many areas, adding difficulty in finding their causal relationship. Existing methods are easy to be over-fitting on them. To fill the gap we mentioned above, in this paper, we propose Shylock, a novel method that can work well in both few-shot and normal MTS to find the causal relationship. Shylock can reduce the number of parameters exponentially by using group dilated convolution and a sharing kernel, but still learn a better representation of variables with time delay. By combing the global constraint and the local constraint, Shylock achieves information sharing among networks to help improve the accuracy. To evaluate the performance of Shylock, we also design a data generation method to generate MTS with time delay. We evaluate it on commonly used benchmarks and generated datasets. Extensive experiments show that Shylock outperforms two existing state-of-art methods on both few-shot and normal MTS. We also developed Tcausal, a library for easy use and deployed it on the EarthDataMiner platform

Problem

• Discovering causal relationships in multivariate time series
• Overcoming data scarcity and overfitting in few-shot scenarios
• Improving accuracy with hybrid constraints and shared representations

Innovation

• Hybrid constraints combine global and local constraints
• Group dilated convolution reduces parameters exponentially
• Shared kernel improves variable representation with time delay