GAT-GAN : A Graph-Attention-based Time-Series Generative Adversarial Network

TL;DR

GAT-GAN introduces a graph-attention-based architecture for generating realistic multivariate time-series data, effectively capturing complex temporal and spatial dependencies, and outperforms existing models on multiple benchmarks.

Contribution

The paper presents a novel GAT-GAN model with graph-attention layers for improved time-series data generation and proposes a new FTD metric for evaluating synthetic data quality.

Findings

GAT-GAN outperforms state-of-the-art benchmarks in fidelity and diversity.

The new FTD metric correlates with downstream predictive performance.

GAT-GAN effectively models long sequences with high fidelity.

Abstract

Generative Adversarial Networks (GANs) have proven to be a powerful tool for generating realistic synthetic data. However, traditional GANs often struggle to capture complex relationships between features which results in generation of unrealistic multivariate time-series data. In this paper, we propose a Graph-Attention-based Generative Adversarial Network (GAT-GAN) that explicitly includes two graph-attention layers, one that learns temporal dependencies while the other captures spatial relationships. Unlike RNN-based GANs that struggle with modeling long sequences of data points, GAT-GAN generates long time-series data of high fidelity using an adversarially trained autoencoder architecture. Our empirical evaluations, using a variety of real-time-series datasets, show that our framework consistently outperforms state-of-the-art benchmarks based on \emph{Frechet Transformer distance} and \emph{Predictive score}, that characterizes (\emph{Fidelity, Diversity}) and \emph{predictive performance} respectively. Moreover, we introduce a Frechet Inception distance-like (FID) metric for time-series data called Frechet Transformer distance (FTD) score (lower is better), to evaluate the quality and variety of generated data. We also found that low FTD scores correspond to the best-performing downstream predictive experiments. Hence, FTD scores can be used as a standardized metric to evaluate synthetic time-series data.