DeepCOVID: An Operational Deep Learning-driven Framework for Explainable Real-time COVID-19 Forecasting

Alexander Rodríguez; Anika Tabassum; Jiaming Cui; Jiajia Xie; Javen Ho; Pulak Agarwal; Bijaya Adhikari; B. Aditya Prakash

DeepCOVID: An Operational Deep Learning-driven Framework for Explainable Real-time COVID-19 Forecasting

Alexander Rodríguez, Anika Tabassum, Jiaming Cui, Jiajia Xie, Javen Ho, Pulak Agarwal, Bijaya Adhikari, B. Aditya Prakash

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

47 Scopus citations

Abstract

How do we forecast an emerging pandemic in real time in a purely data-driven manner? How to leverage rich heterogeneous data based on various signals such as mobility, testing, and/or disease exposure for forecasting? How to handle noisy data and generate uncertainties in the forecast? In this paper, we present DEEPCOVID, an operational deep learning framework designed for real-time COVID-19 forecasting. DEEPCOVID works well with sparse data and can handle noisy heterogeneous data signals by propagating the uncertainty from the data in a principled manner resulting in meaningful uncertainties in the forecast. The deployed framework also consists of modules for both real-time and retrospective exploratory analysis to enable interpretation of the forecasts. Results from real-time predictions (featured on the CDC website and FiveThirtyEight.com) since April 2020 indicates that our approach is competitive among the methods in the COVID-19 Forecast Hub, especially for short-term predictions.

Original language	English
Title of host publication	35th AAAI Conference on Artificial Intelligence, AAAI 2021
Publisher	Association for the Advancement of Artificial Intelligence
Pages	15393-15400
Number of pages	8
ISBN (Electronic)	9781713835974
State	Published - 2021
Externally published	Yes
Event	35th AAAI Conference on Artificial Intelligence, AAAI 2021 - Virtual, Online Duration: Feb 2 2021 → Feb 9 2021

Publication series

Name	35th AAAI Conference on Artificial Intelligence, AAAI 2021
Volume	17B

Conference

Conference	35th AAAI Conference on Artificial Intelligence, AAAI 2021
City	Virtual, Online
Period	02/2/21 → 02/9/21

Funding

We would like to thank the anonymous reviewers for their helpful suggestions which improved the paper. This paper is based on work partially supported by the NSF (Expeditions CCF-1918770, CAREER IIS-2028586, RAPID IIS-2027862, Medium IIS-1955883, NRT DGE-1545362), CDC MInD program, ORNL and funds/computing resources from Georgia Tech and GTRI. B. A. was in part supported by the CDC MInD-Healthcare U01CK000531-Supplement.

Cite this

Rodríguez, A., Tabassum, A., Cui, J., Xie, J., Ho, J., Agarwal, P., Adhikari, B., & Prakash, B. A. (2021). DeepCOVID: An Operational Deep Learning-driven Framework for Explainable Real-time COVID-19 Forecasting. In 35th AAAI Conference on Artificial Intelligence, AAAI 2021 (pp. 15393-15400). (35th AAAI Conference on Artificial Intelligence, AAAI 2021; Vol. 17B). Association for the Advancement of Artificial Intelligence.

Rodríguez, Alexander ; Tabassum, Anika ; Cui, Jiaming et al. / DeepCOVID : An Operational Deep Learning-driven Framework for Explainable Real-time COVID-19 Forecasting. 35th AAAI Conference on Artificial Intelligence, AAAI 2021. Association for the Advancement of Artificial Intelligence, 2021. pp. 15393-15400 (35th AAAI Conference on Artificial Intelligence, AAAI 2021).

@inproceedings{1c4eabd9233e46c0855383dcb6db2cfa,

title = "DeepCOVID: An Operational Deep Learning-driven Framework for Explainable Real-time COVID-19 Forecasting",

abstract = "How do we forecast an emerging pandemic in real time in a purely data-driven manner? How to leverage rich heterogeneous data based on various signals such as mobility, testing, and/or disease exposure for forecasting? How to handle noisy data and generate uncertainties in the forecast? In this paper, we present DEEPCOVID, an operational deep learning framework designed for real-time COVID-19 forecasting. DEEPCOVID works well with sparse data and can handle noisy heterogeneous data signals by propagating the uncertainty from the data in a principled manner resulting in meaningful uncertainties in the forecast. The deployed framework also consists of modules for both real-time and retrospective exploratory analysis to enable interpretation of the forecasts. Results from real-time predictions (featured on the CDC website and FiveThirtyEight.com) since April 2020 indicates that our approach is competitive among the methods in the COVID-19 Forecast Hub, especially for short-term predictions.",

author = "Alexander Rodr{\'i}guez and Anika Tabassum and Jiaming Cui and Jiajia Xie and Javen Ho and Pulak Agarwal and Bijaya Adhikari and Prakash, \{B. Aditya\}",

note = "Publisher Copyright: Copyright {\textcopyright} 2021, Association for the Advancement of Artificial Intelligence (www.aaai.org). All rights reserved; 35th AAAI Conference on Artificial Intelligence, AAAI 2021 ; Conference date: 02-02-2021 Through 09-02-2021",

year = "2021",

language = "English",

series = "35th AAAI Conference on Artificial Intelligence, AAAI 2021",

publisher = "Association for the Advancement of Artificial Intelligence",

pages = "15393--15400",

booktitle = "35th AAAI Conference on Artificial Intelligence, AAAI 2021",

}

Rodríguez, A, Tabassum, A, Cui, J, Xie, J, Ho, J, Agarwal, P, Adhikari, B & Prakash, BA 2021, DeepCOVID: An Operational Deep Learning-driven Framework for Explainable Real-time COVID-19 Forecasting. in 35th AAAI Conference on Artificial Intelligence, AAAI 2021. 35th AAAI Conference on Artificial Intelligence, AAAI 2021, vol. 17B, Association for the Advancement of Artificial Intelligence, pp. 15393-15400, 35th AAAI Conference on Artificial Intelligence, AAAI 2021, Virtual, Online, 02/2/21.

DeepCOVID: An Operational Deep Learning-driven Framework for Explainable Real-time COVID-19 Forecasting. / Rodríguez, Alexander; Tabassum, Anika; Cui, Jiaming et al.
35th AAAI Conference on Artificial Intelligence, AAAI 2021. Association for the Advancement of Artificial Intelligence, 2021. p. 15393-15400 (35th AAAI Conference on Artificial Intelligence, AAAI 2021; Vol. 17B).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - DeepCOVID

T2 - 35th AAAI Conference on Artificial Intelligence, AAAI 2021

AU - Rodríguez, Alexander

AU - Tabassum, Anika

AU - Cui, Jiaming

AU - Xie, Jiajia

AU - Ho, Javen

AU - Agarwal, Pulak

AU - Adhikari, Bijaya

AU - Prakash, B. Aditya

PY - 2021

Y1 - 2021

N2 - How do we forecast an emerging pandemic in real time in a purely data-driven manner? How to leverage rich heterogeneous data based on various signals such as mobility, testing, and/or disease exposure for forecasting? How to handle noisy data and generate uncertainties in the forecast? In this paper, we present DEEPCOVID, an operational deep learning framework designed for real-time COVID-19 forecasting. DEEPCOVID works well with sparse data and can handle noisy heterogeneous data signals by propagating the uncertainty from the data in a principled manner resulting in meaningful uncertainties in the forecast. The deployed framework also consists of modules for both real-time and retrospective exploratory analysis to enable interpretation of the forecasts. Results from real-time predictions (featured on the CDC website and FiveThirtyEight.com) since April 2020 indicates that our approach is competitive among the methods in the COVID-19 Forecast Hub, especially for short-term predictions.

AB - How do we forecast an emerging pandemic in real time in a purely data-driven manner? How to leverage rich heterogeneous data based on various signals such as mobility, testing, and/or disease exposure for forecasting? How to handle noisy data and generate uncertainties in the forecast? In this paper, we present DEEPCOVID, an operational deep learning framework designed for real-time COVID-19 forecasting. DEEPCOVID works well with sparse data and can handle noisy heterogeneous data signals by propagating the uncertainty from the data in a principled manner resulting in meaningful uncertainties in the forecast. The deployed framework also consists of modules for both real-time and retrospective exploratory analysis to enable interpretation of the forecasts. Results from real-time predictions (featured on the CDC website and FiveThirtyEight.com) since April 2020 indicates that our approach is competitive among the methods in the COVID-19 Forecast Hub, especially for short-term predictions.

UR - http://www.scopus.com/inward/record.url?scp=85102342229&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:85102342229

T3 - 35th AAAI Conference on Artificial Intelligence, AAAI 2021

SP - 15393

EP - 15400

BT - 35th AAAI Conference on Artificial Intelligence, AAAI 2021

PB - Association for the Advancement of Artificial Intelligence

Y2 - 2 February 2021 through 9 February 2021

ER -

Rodríguez A, Tabassum A, Cui J, Xie J, Ho J, Agarwal P et al. DeepCOVID: An Operational Deep Learning-driven Framework for Explainable Real-time COVID-19 Forecasting. In 35th AAAI Conference on Artificial Intelligence, AAAI 2021. Association for the Advancement of Artificial Intelligence. 2021. p. 15393-15400. (35th AAAI Conference on Artificial Intelligence, AAAI 2021).

DeepCOVID: An Operational Deep Learning-driven Framework for Explainable Real-time COVID-19 Forecasting

Abstract

Publication series

Conference

Funding

Other files and links

Fingerprint

Cite this