Role of Machine Learning in fighting COVID-19
As of April 1st 2020, the number of recorded cases has reached around 900,000 with no end at sight.
COVID-19 is an infectious disease caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). The disease was first identified in December 2019 in Wuhan, the capital of China’s Hubei province, and has since spread globally, resulting in the ongoing 2019–20 coronavirus pandemic. As of April 1st 2020, the number of recorded cases has reached around 900,000 with no end at sight. In case of such peril, every researcher wants to help out in fighting the disease with what little they can do. The field of machine learning and data analytics is not far from playing its part either.
In this article, I will be addressing the different categories where Machine Learning has played or could potentially play an essential role. Please note that I am not a trained specialist in fields like epidemiology, chemistry or biology but as a Machine Learning professional, I am trained to be flexible and apply my knowledge in any possible field.
According to [1], the applications can be classified into the following specialized fields:
- Clinical diagnostics
- Forecasting of disease progression
- Development of anti-viral drugs
- Public sentiment analysis
- Efficient data augmentation and utilization
Clinical diagnostics
In the work by Wang et al [8], they developed an Inception transfer-learning based model, followed by internal and external validation that could extract COVID-19’s graphical features in order to provide a clinical diagnosis ahead of the pathogenic test, so as to save critical time for disease control. The method was able to extract radiological features for timely and accurate COVID-19 diagnosis with 85.2% accuracy.
The clinical research by Chen et al [9], implemented UNet++, a novel and powerful architecture for medical image segmentation for detecting 2019 novel coronavirus pneumonia on 2 high-resolution computed tomography. The model showed comparable performance with expert radiologists and greatly improve the efficiency of radiologists in clinical practice. It holds great potential to relieve the pressure off frontline radiologists, improve early diagnosis, isolation and treatment, and thus contribute to the control of the epidemic.
The work of Xu et al [10] aimed to establish an early screening model to distinguish COVID-19 pneumonia from Influenza-A viral pneumonia and healthy cases with pulmonary CT images using two CNN three-dimensional classification models — ResNet-18 and other with an additional location-attention mechanism in the full-connection layer.
Forecasting of disease progression
According to the paper by Yan et al [6], they performed a clinical study on patients admitted in Tongji Hospital from January 10th to February 18th, 2020 and developed a prognostic prediction model based on XGBoost machine learning algorithm. Using this, they were able to predict the mortality risk, and present a clinical route to the recognition of critical cases from severe cases which could help doctors with early identification and intervention, thus potentially reducing mortality.
Anastassopoulou et al [7] presented a study where they used Susceptible-Infected-Recovered-Dead (SIRD) model to provide tentative three-week forecasts of the evolution of the outbreak at the epicentre, Wuhan. The analysis further revealed a significant decline of the mortality rate to which various factors may have contributed, such as the severe control measures taken in Hubei, China (e.g. quarantine and hospitalization of infected individuals).
Development of anti-viral drugs
The process of drug development plays a vital role in preventing, controlling and treating any form of the disease. The research that is being worked on so far predict the effectiveness of present drugs and treatment procedures through deep networks. According to Zhang et al [3], as the drug development process requires a long period, alternative methods are the need of the hour. In his paper, coronavirus main protease (considered to be a major therapeutic target) has been the focus of drug screening process using the DFCNN which identifies the protein-ligand interactions with high accuracy.
Another study by Beck et al [4] used pre-trained deep learning-based drug-target interaction model called Molecule Transformer-Drug Target Interaction (MT-DTI) to identify commercially available drugs that could act on viral proteins of 2019-nCoV. They suggested a list of antiviral drugs identified by the MT-DTI model that could be considered when establishing effective treatment strategies for 2019-nCoV. Similarly, DeepMind [5] has published computation predictions of protein structure using their AlphaFold system.
Public Sentiment Analysis
Early detection and prediction of a disease outbreak is critical because it can provide more time to prepare a response and significantly reduce the impact caused by a pandemic. Alessa et al [11] worked on tracking trends of Influenza-Like Illness (ILI) from search engines and social networking sites so as to identify it seven to ten days in advance using various methods like support vector machines and neural networks. They provided an insightful review of Influenza virus spread which could be used as a reference for COVID-19 tracking. Similar work has been presented by Lee et al [12] on Centers for Disease Control and Prevention (CDC) data.
The paper by Wang et al [13] presented a method for influenza prediction based on the real-time geotagged tweet data from social media so as to ensures real-time prediction and is applicable to sampling data. They used partial differential equations (PDEs) model which incorporated factors like effects of flu spreading, flu recovery, and active human interventions for reducing flu.
Efficient data augmentation and utilization
The outpouring of research is a testament to the speed with which science can tackle big problems. But it also presents a challenge in scanning literature to obtain insights. Efforts can be made to mine through the avalanche of research to answer questions that could help medical and public health experts. The sector has huge potential to help wrangle and draw insights from scientific research [14]. But in my opinion, the approach is at an early stage and is unlikely to help address the current crisis.
This wraps up my take on the role of machine learning in fighting COVD-19. If you made it till here, thank you for supporting my work. Please note that I am open to critique from scholars and experts from relevant fields.
[1] Godfried, I., 2020. Machine Learning Methods To Aid In Coronavirus Response. [online] Medium. Available at: https://towardsdatascience.com/machine-learning-methods-to-aid-in-coronavirus-response-70df8bfc7861 [Accessed 2 April 2020].
[2] World Economic Forum. 2020. This Interactive Chart Shows Countries Flattening Their COVID-19 Curves. [online] Available at: https://www.weforum.org/agenda/2020/03/covid19-coronavirus-countries-infection-trajectory/ [Accessed 2 April 2020].
[3] Zhang, H., Saravanan, K.M., Yang, Y., Hossain, M.T., Li, J., Ren, X. and Wei, Y., 2020. Deep learning based drug screening for novel coronavirus 2019-nCov.
[4] Beck, B.R., Shin, B., Choi, Y., Park, S. and Kang, K., 2020. Predicting commercially available antiviral drugs that may act on the novel coronavirus (2019-nCoV), Wuhan, China through a drug-target interaction deep learning model. bioRxiv.
[5] Deepmind. 2020. Computational Predictions Of Protein Structures Associated With COVID-19. [online] Available at: https://deepmind.com/research/open-source/computational-predictions-of-protein-structures-associated-with-COVID-19 [Accessed 2 April 2020].
[6] Yan, L., Zhang, H.T., Xiao, Y., Wang, M., Sun, C., Liang, J., Li, S., Zhang, M., Guo, Y., Xiao, Y. and Tang, X., 2020. Prediction of criticality in patients with severe Covid-19 infection using three clinical features: a machine learning-based prognostic model with clinical data in Wuhan. medRxiv.
[7] Anastassopoulou, C., Russo, L., Tsakris, A. and Siettos, C., 2020. Data-Based Analysis, Modelling and Forecasting of the novel Coronavirus (2019-nCoV) outbreak. medRxiv.
[8] Wang, S., Kang, B., Ma, J., Zeng, X., Xiao, M., Guo, J., Cai, M., Yang, J., Li, Y., Meng, X. and Xu, B., 2020. A deep learning algorithm using CT images to screen for Corona Virus Disease (COVID-19). medRxiv.
[9] Chen, J., Wu, L., Zhang, J., Zhang, L., Gong, D., Zhao, Y., Hu, S., Wang, Y., Hu, X., Zheng, B. and Zhang, K., 2020. Deep learning-based model for detecting 2019 novel coronavirus pneumonia on high-resolution computed tomography: a prospective study. medRxiv.
[10] Xu, X., Jiang, X., Ma, C., Du, P., Li, X., Lv, S., Yu, L., Chen, Y., Su, J., Lang, G. and Li, Y., 2020. Deep Learning System to Screen Coronavirus Disease 2019 Pneumonia. arXiv preprint arXiv:2002.09334.
[11] Alessa, A. and Faezipour, M., 2018. A review of influenza detection and prediction through social networking sites. Theoretical Biology and Medical Modelling, 15(1), p.2.
[12] Lee, K., Agrawal, A. and Choudhary, A., 2017, August. Forecasting influenza levels using real-time social media streams. In 2017 IEEE International Conference on Healthcare Informatics (ICHI) (pp. 409–414). IEEE.
[13] Wang, Y., Xu, K., Kang, Y., Wang, H., Wang, F. and Avram, A., 2020. Regional Influenza Prediction with Sampling Twitter Data and PDE Model. International journal of environmental research and public health, 17(3), p.678.
[14] Knight, W., 2020. Researchers Will Deploy AI To Better Understand Coronavirus. [online] Wired. Available at: https://www.wired.com/story/researchers-deploy-ai-better-understand-coronavirus/ [Accessed 2 April 2020].
