Inter-Series Attention Model for COVID-19 Forecasting

TL;DR

This paper introduces ACTS, a neural network model using inter-series attention to forecast COVID-19 trends across regions, outperforming traditional compartmental models in most tests.

Contribution

The paper proposes a novel attention-based neural model for COVID-19 forecasting that leverages shared progression patterns across regions, surpassing existing models in accuracy.

Findings

ACTS outperforms CDC-referenced models in 13 out of 18 tests.

The model effectively captures shared progression patterns across different regions.

Attention mechanisms improve the accuracy of disease trend forecasting.

Abstract

COVID-19 pandemic has an unprecedented impact all over the world since early 2020. During this public health crisis, reliable forecasting of the disease becomes critical for resource allocation and administrative planning. The results from compartmental models such as SIR and SEIR are popularly referred by CDC and news media. With more and more COVID-19 data becoming available, we examine the following question: Can a direct data-driven approach without modeling the disease spreading dynamics outperform the well referred compartmental models and their variants? In this paper, we show the possibility. It is observed that as COVID-19 spreads at different speed and scale in different geographic regions, it is highly likely that similar progression patterns are shared among these regions within different time periods. This intuition lead us to develop a new neural forecasting model, called Attention Crossing Time Series (\textbf{ACTS}), that makes forecasts via comparing patterns across time series obtained from multiple regions. The attention mechanism originally developed for natural language processing can be leveraged and generalized to materialize this idea. Among 13 out of 18 testings including forecasting newly confirmed cases, hospitalizations and deaths, \textbf{ACTS} outperforms all the leading COVID-19 forecasters highlighted by CDC.