Abstract:Graph neural networks (GNNs), especially dynamic GNNs, have become a research hotspot in spatio-temporal forecasting problems. While many dynamic graph construction methods have been developed, relatively few of them explore the causal relationship between neighbour nodes. Thus, the resulting models lack strong explainability for the causal relationship between the neighbour nodes of the dynamically generated graphs, which can easily lead to a risk in subsequent decisions. Moreover, few of them consider the uncertainty and noise of dynamic graphs based on the time series datasets, which are ubiquitous in real-world graph structure networks. In this paper, we propose a novel Dynamic Diffusion-Variational Graph Neural Network (DVGNN) for spatio-temporal forecasting. For dynamic graph construction, an unsupervised generative model is devised. Two layers of graph convolutional network (GCN) are applied to calculate the posterior distribution of the latent node embeddings in the encoder stage. Then, a diffusion model is used to infer the dynamic link probability and reconstruct causal graphs in the decoder stage adaptively. The new loss function is derived theoretically, and the reparameterization trick is adopted in estimating the probability distribution of the dynamic graphs by Evidence Lower Bound during the backpropagation period. After obtaining the generated graphs, dynamic GCN and temporal attention are applied to predict future states. Experiments are conducted on four real-world datasets of different graph structures in different domains. The results demonstrate that the proposed DVGNN model outperforms state-of-the-art approaches and achieves outstanding Root Mean Squared Error result while exhibiting higher robustness. Also, by F1-score and probability distribution analysis, we demonstrate that DVGNN better reflects the causal relationship and uncertainty of dynamic graphs.

What problem does this paper attempt to address?

The paper primarily aims to address two key challenges in spatiotemporal prediction: causal interpretability and uncertainty handling in dynamic graph structures. ### Overview of the Problems Addressed by the Paper 1. **Causal Interpretability**: Existing dynamic graph neural network methods often overlook the causal relationships between nodes when constructing dynamic graphs, resulting in graphs that lack strong interpretability. This can pose risks in subsequent decision-making. 2. **Uncertainty and Noise Handling**: Most methods do not adequately consider the uncertainty and noise in dynamic graphs based on time series data, which is a common issue in real-world graph structure networks. ### Proposed Method To tackle the aforementioned challenges, the authors propose a new model called **Dynamic Variational Graph Neural Network (DVGNN)**. This model combines Graph Convolutional Networks (GCN) and diffusion processes to enhance the accuracy, robustness, and interpretability of spatiotemporal predictions. - **Model Architecture**: DVGNN includes an encoder phase and a decoder phase. The encoder phase uses two layers of GCN to compute the posterior distribution of latent node embeddings; the decoder phase employs a diffusion model to infer dynamic link probabilities and reconstruct causal graphs. - **Causal Relationship Exploration**: Through the diffusion process, the model can infer the intrinsic causal relationships between adjacent nodes from the relationships between latent states. - **Uncertainty Modeling**: By using noise-driven Stochastic Differential Equations (SDE), the model captures the uncertainty in spatiotemporal data, thereby better capturing the complexity and variability in dynamic graphs. - **Experimental Validation**: The study conducted experiments on real-world datasets from four different domains, demonstrating the superior performance of DVGNN in spatiotemporal prediction tasks. It also showcased the model's advantages in predicting the uncertainty of dynamic edges, robustness, and the interpretability of generated graphs. In summary, the paper addresses the key challenges faced by existing methods in handling spatiotemporal prediction problems, particularly in terms of causal interpretability and uncertainty handling in dynamic graph structures, by proposing DVGNN.

Dynamic Causal Explanation Based Diffusion-Variational Graph Neural Network for Spatio-temporal Forecasting

Dynamic Causal Explanation Based Diffusion-Variational Graph Neural Network for Spatiotemporal Forecasting

Dynamic Spatio-Temporal Graph Network with Adaptive Propagation Mechanism for Multivariate Time Series Forecasting

Dynamic Personalized Graph Neural Network with Linear Complexity for Multivariate Time Series Forecasting

When Dynamic Causality Comes to Graph-Temporal Neural Network

Dynamic Graph Learning-Neural Network for Multivariate Time Series Modeling

CausalGNN: Causal-Based Graph Neural Networks for Spatio-Temporal Epidemic Forecasting

Spatiotemporal Forecasting Meets Efficiency: Causal Graph Process Neural Networks

Dynamic Causal Graph Convolutional Network for Traffic Prediction

MDG: A Multi-Task Dynamic Graph Generation Framework for Multivariate Time Series Forecasting

GDBN: a Graph Neural Network Approach to Dynamic Bayesian Network

Time-aware Personalized Graph Convolutional Network for Multivariate Time Series Forecasting

Dynamic Graph Structure Learning for Multivariate Time Series Forecasting

TimeGNN: Temporal Dynamic Graph Learning for Time Series Forecasting

Multiple Time Series Forecasting with Dynamic Graph Modeling

Dynamic Spatial-Temporal Graph Convolutional Neural Networks for Traffic Forecasting.

Spatial dynamic graph convolutional network for traffic flow forecasting

Connecting the Dots: Multivariate Time Series Forecasting with Graph Neural Networks

Dynamic Adaptive and Adversarial Graph Convolutional Network for Traffic Forecasting

Multivariate Time Series Forecasting with Transfer Entropy Graph

Scalable Spatiotemporal Graph Neural Networks