COVID-19 has affected the worldwide economy, politics, education, tourism, and actually EVERYTHING. Many academic papers address trends prediction in various fields due to COVID-19, with the power of Artificial Intelligence.
COVID-19 pandemic has affected the entire world. Many people lost their jobs, kids stay at home, and the economic crisis is disastrous. The question of “how will the world be after COVID-19” is of high interest. Many futurists predict a different world, where we should rethink public spaces and believes that the memory of the COVID-19 lockdown will remain for a long time (Del Bello, 2020). Google collects and arranges worldwide data, as you can see the daily new cases and deaths:
This information presents a sad situation, where the COVID-19 continues to spreads with tragic death cases. Aside from Google, many other data sources are available online in the era of big data.
Artificial Intelligence (AI) approaches have been involved to recognize and predict new cases. The power of AI allows more than that, by design deep learning-based- models, one can predict worldwide trends. In this post, we present the COVID-Net that allows case detection by X-ray images. Then, we review how Google Trends helps in prediction tasks, how satellite images predict economic crises and quantify their trends, tourism “back to normal” prediction, and how managing COVID-19 lockdown by GPS users localization. All presented approaches are AI-based methods, mostly deep learning.
In May 2020, three Canadian researchers published a novel deep learning model for the detection of COVID-19 cases from chest X-ray images (Wang et al.). They generated the COVIDx dataset from available open access data repositories that included almost 14,000 images. The COVID-Net architecture provides enhanced capacity while keeping computational efficiency:
Their model was compared to the VGG-19 and the ResNet-50 and overcame them with higher accuracy:
The COVID-Net leverages area in the lungs at the main critical factors in the classification task. By that, they use an explainability-driven auditing strategy and improve the reliability of deep learning for clinical applications.
Search engines provide useful data that might be useful to analyze COVID-19 trends. Google Trends provides a measure entitled interest over time (between 0 to 100). A value of 50 means that the term you insert is “half popular”. Recently, Ayyoubzadeh et al. predicted COVID-19 trends in Iran by using the following features set:
These features are input to Long Short-Term Memory (LSTM) architecture and logistic regression. Their LSTM model predicted new cases by considering small windows. The regression seemed to work better.
Generally, Google Trends seems to be a great tool for data scientists. For example, we compared the terms “Flight” and “Covid19”, where May 22, 2020, is a shifting point.
COVID19 affects the global economy. The effect is so huge, that it leaves “signatures” in satellite images. For example, in the figures below we can see the” before” and “after” for airports and car rental parking lot. This crisis can be observed in a precise way by an analysis and detection system that can observe the relevant elements in images and counts for the relevant objects. Recently, Minetto et al suggested using AI for this task.
The AI system allows mobility detection during COVID-19. By analyzing the mobility worldwide, many economic trends can be detected and the “rule-nakers” can use this information. The model flowchart is given below, where in the first step, the area of interest is chosen with location sampling (a)-(b), then, location boundaries are set according to roads and street (c), an image processing stages include region-of-interest extraction, resizing, splitting, detection, filtering, and merging (d)-(g).
One key priority is to prevent the epidemic spread. Many countries are in lockdown and ensure the social distancing of their population. As manual contact-tracing is highly inefficient and time-consuming, a smartphone-based approach to automatically manage the lockdown by widely tracing the contacts by GPS was suggested by Maghdid et al. Creating a list of individuals allows the notification of confirmed COVID-19 cases.
The prediction model uses the K-means algorithm (unsupervised learning technique) for lockdown management, it obtains user’s region and position and returns the centroid of all positions assigned to the cluster. By that, the nearest of each user to a confirmed case can be obtained.
This approach allows the decision-makers to consider real-time information in efficiently. As more users join the application, more data be collected and the model predicts better.
COVID-19 hardly affects tourism. Recently, Polyzos et al suggest predicting “back-to-normal” tourism by comparing it to SARS using the LSTM model. They focus on Chines arrivals to the USA and Australia. The model predicts a period of about half a year to one year to return to full activity, from the moment all restrictions will be removed.
The power of Artificial Intelligence is enormous, where it has first been used during a worldwide pandemic, COVID-19. We present some academic papers dealing with prediction and tasks that help with pandemic management, where they used as a tool for the decision-makers.
— — — — — — — — — — — — — — — — — —
Barak Or received the B.Sc. (2016), M.Sc. (2018) degrees in aerospace engineering, and also B.A. in economics and management (2016, Cum Laude) from the Technion, Israel Institute of Technology. He was with Qualcomm (2019–2020), where he mainly dealt with Machine Learning and Signal Processing algorithms. Barak currently studies toward his Ph.D. at the University of Haifa. His research interest includes sensor fusion, navigation, deep learning, and estimation theory.
Visit my personal website: www.Barakor.com
— — — — — — — — — — — — — — — — –
 Lou Del Bello. “How Covid-19 could redesign our world”. BBC Future, 28th May 2020.
 Wang, Linda, and Alexander Wong. “COVID-Net: A Tailored Deep Convolutional Neural Network Design for Detection of COVID-19 Cases from Chest X-Ray Images.” arXiv preprint arXiv:2003.09871 (2020).
 Ayyoubzadeh, Seyed Mohammad, et al. “Predicting COVID-19 incidence through analysis of google trends data in iran: data mining and deep learning pilot study.” JMIR Public Health and Surveillance 6.2 (2020): e18828.
 Minetto, Rodrigo, et al. “Measuring Human and Economic Activity from Satellite Imagery to Support City-Scale Decision-Making during COVID-19 Pandemic.” arXiv preprint arXiv:2004.07438 (2020).
 Maghdid, Halgurd S., and Kayhan Zrar Ghafoor. “A smartphone enabled approach to manage COVID-19 lockdown and economic crisis.” arXiv preprint arXiv:2004.12240 (2020).
 Polyzos, Stathis, Aristeidis Samitas, and Anastasia Ef Spyridou. “Tourism demand and the covid-19 pandemic: an lstm approach.” Tourism Recreation Research (2020): 1–13.