Abstract:Causality analysis is an important problem lying at the heart of science, and is of particular importance in data science and machine learning. An endeavor during the past 16 years viewing causality as real physical notion so as to formulate it from first principles, however, seems to go unnoticed. This study introduces to the community this line of work, with a long-due generalization of the information flow-based bivariate time series causal inference to multivariate series, based on the recent advance in theoretical development. The resulting formula is transparent, and can be implemented as a computationally very efficient algorithm for application. It can be normalized, and tested for statistical significance. Different from the previous work along this line where only information flows are estimated, here an algorithm is also implemented to quantify the influence of a unit to itself. While this forms a challenge in some causal inferences, here it comes naturally, and hence the identification of self-loops in a causal graph is fulfilled automatically as the causalities along edges are inferred. To demonstrate the power of the approach, presented here are two applications in extreme situations. The first is a network of multivariate processes buried in heavy noises (with the noise-to-signal ratio exceeding 100), and the second a network with nearly synchronized chaotic oscillators. In both graphs, confounding processes exist. While it seems to be a huge challenge to reconstruct from given series these causal graphs, an easy application of the algorithm immediately reveals the desideratum. Particularly, the confounding processes have been accurately differentiated. Considering the surge of interest in the community, this study is very timely.

Bi-directional Causal Graph Learning Through Weight-Sharing and Low-Rank Neural Network

Learning Cluster Causal Diagrams: an Information-Theoretic Approach

CausalLP: Learning causal relations with weighted knowledge graph link prediction

Towards the Disappearing Truth: Fine-Grained Joint Causal Influences Learning with Hidden Variable-Driven Causal Hypergraphs in Time Series

TS-CausalNN: Learning Temporal Causal Relations from Non-linear Non-stationary Time Series Data

Rethinking Causal Relationships Learning in Graph Neural Networks

Neural Network Weight Comparison for Industrial Causality Discovering and Its Soft Sensing Application

Normalized multivariate time series causality analysis and causal graph reconstruction

Invariant Graph Learning for Causal Effect Estimation

Causality-based CTR Prediction using Graph Neural Networks

CUTS+: High-dimensional Causal Discovery from Irregular Time-series

Deep Learning-based Group Causal Inference in Multivariate Time-series

Graph Neural Network Causal Explanation via Neural Causal Models

On Low Rank Directed Acyclic Graphs and Causal Structure Learning

Neural Causal Graph Collaborative Filtering

GDBN: a Graph Neural Network Approach to Dynamic Bayesian Network

CUTS: Neural Causal Discovery from Unstructured Time-Series Data

CUTS: Neural Causal Discovery from Irregular Time-Series Data

De Bruijn goes Neural: Causality-Aware Graph Neural Networks for Time Series Data on Dynamic Graphs

Directed Cyclic Graph for Causal Discovery from Multivariate Functional Data

Amortized Causal Discovery: Learning to Infer Causal Graphs from Time-Series Data