Asynchronous Graph Generator

TL;DR

The paper presents AGG, a novel attention-based graph neural network for multi-channel time series imputation and prediction that outperforms existing methods without relying on recurrent structures.

Contribution

Introduces the asynchronous graph generator (AGG), a new graph attention network that encodes multi-channel time series data for improved imputation and prediction.

Findings

Achieved state-of-the-art results on benchmark datasets.

Outperformed recent attention-based networks.

Effective in data augmentation scenarios.

Abstract

We introduce the asynchronous graph generator (AGG), a novel graph attention network for imputation and prediction of multi-channel time series. Free from recurrent components or assumptions about temporal/spatial regularity, AGG encodes measurements, timestamps and channel-specific features directly in the nodes via learnable embeddings. Through an attention mechanism, these embeddings allow for discovering expressive relationships among the variables of interest in the form of a homogeneous graph. Once trained, AGG performs imputation by \emph{conditional attention generation}, i.e., by creating a new node conditioned on given timestamps and channel specification. The proposed AGG is compared to related methods in the literature and its performance is analysed from a data augmentation perspective. Our experiments reveal that AGG achieved state-of-the-art results in time series imputation, classification and prediction for the benchmark datasets \emph{Beijing Air Quality}, \emph{PhysioNet ICU 2012} and \emph{UCI localisation}, outperforming other recent attention-based networks.