AI and IoT enabled framework for automatic pandemic management
When countries around the world witnessed an early exponential increase in COVID-19 cases in their towns, the world panicked. This has led them to respond in various ways, such as formulating new policies to deal with the COVID-19 pandemic, combining the use of advanced technologies and competition in drug and vaccine development.
Lockdown is a universal policy, most countries follow to suspend the volume of cases and prepare to manage the exponential growth of COVID-19 cases in the near future. Countries/regions such as Taiwan, South Korea, Singapore, and New Zealand were able to effectively manage and control COVID-19 without panic. They all used testing as a powerful tool. On the other hand, larger countries, although later, saw extreme panics, such as Brazil, India, Russia, the United Kingdom, and the United States.
This shows policy failure at various levels because these countries had been alerted beforehand. Chain breaking showed little progress, but despite this, they have not made result-oriented decisions. Another mistake made by some countries is to use lockdown as the main tool to control COVID-19 (in India, for example) instead of focusing on testing and improving medical infrastructure to the maximum.
Effective management of a pandemic requires a multifaceted effort and policy formulation, including planning, advanced technology, data sharing, financial resources, logistics, and transparency.
Literature review
This article focuses on the technical aspects of pandemic management, with particular emphasis on integrating various technologies to achieve a common goal. In this section, we discuss research and development work that focuses on manufacturing AI-based sensors and technologies to control the spread of infectious diseases.
The COVID-19 pandemic started with a local infection in Wuhan and has now infected more than 166 countries around the world. As we have seen today, artificial intelligence-based technologies have been successfully applied in various fields to improve decision-making accuracy and achieve complete automation.
Nowadays, machine learning and deep learning have become the core of any AI-based technology and tools, such as big data analysis, social network analysis, network, autonomous driving, biology, Health, Astronomy, Physics, and Transportation, are examples.
Sentiment analysis can play an important role in formulating policies to deal with COVID-19. Social media and online news are unstructured data sources that contain opinions and information about the behavior of individuals, communities, and events. One of the powerful tools that can help control the spread of COVID-19 is to provide the public with correct information about what to do and don’t do during a pandemic.
Today, the world has witnessed an information epidemic where there is too much correct and incorrect information (for example, what type of masks to use, magic drugs, treatment methods, rules to follow, and isolation guidelines). Sentiment analysis can help find solutions to the COVID-19 information epidemic.
A study used machine learning to analyze the development of fear in the public through tweets and found that as the number of cases increases, fear gradually increases. However, people are now showing signs of fatigue. Similarly, in another study, Nemes and Kiss used machine learning to find interesting trends related to COVID-19 on Twitter. In this trend, a chatbot tool based on sentiment analysis uses deep learning to interact with quarantined patients. To keep them fresh and stay away from negative thoughts.
Social distancing is the key to breaking the chain of infection and the ability to continue spreading. Analysis of social networks and recognition of human activity can help track people and events. People’s emotions can also be predicted using unstructured social media data. However, there will be some privacy issues that can be resolved individually.
Medical image segmentation is widely used in infectious disease detection. For example, a deep learning-based method was introduced to use their CT images to identify patients infected with COVID-19. However, these methods can improve their prediction accuracy. One of the reasons for the low accuracy is limited data availability.
Chen et al. proposed a detection accuracy improvement method based on deep learning using high-resolution computed tomography; for scientists, deep learning plays an important role in improving their understanding of COVID-19.
On the other hand, drug discovery plays a very important role in the control of highly infectious diseases such as COVID-19. Drug development, especially for newly discovered pathogens, is a time-consuming task.
Several benefits of machine learning and deep learning can be used to design drugs, such as identifying target populations, identifying prognostic biomarkers, examining pathological data sets, and data-driven decision-making. Chen et al. compare deep learning with traditional machine learning algorithms for drug discovery problems; deep learning outperforms all other knowledge. Next, a deep learning-based method was proposed to classify protein-ligand interactions. This method can identify drug candidates for COVID-19 protease by screening drugs from four compound databases.
In order to effectively manage and control COVID-19, some countries have used Internet of Things (IoT) applications integrated with AI for monitoring and contact tracking. The Internet of Things and artificial intelligence has been widely used in the healthcare industry, such as IBM Watson, which can find valuable information from unstructured data and can even correct doctors based on their prescriptions and diagnoses.
During an infectious disease outbreak, artificial intelligence can answer several complex questions. One of these applications is BlueDot. It is China’s first AI-based tool to quantify the early global risk of COVID-19. He communicated the risk factors to the customer. Use a wide variety of data sets to predict the impact of infections, such as global flight routes, weather conditions, animal and insect populations, and the capabilities of specific regional health systems.
Proposed Framework
The key to dealing with the highly infectious COVID-19 is to have an efficient response system that is intelligent, automated, and capable of making decisions quickly. However, as we have seen in recent months, even developed countries in the world have not been able to establish any state-of-the-art response systems on a large scale. From a technical point of view, the main reason for this: the lack of integration between the various subsystems. This also means poor coordination.
For example, AI-enabled tools can predict the outbreak of COVID-19; however, several countries do not have requirements for preparing medical equipment and medical supplies based on threats. Once the number of cases multiplied, they rushed to obtain critical medical needs. Especially in Italy, the United States, Brazil, and many other countries, this happened between February and April 2020. At each stage, there will be a lot of human intervention, which slows down the decision-making process and is full of personal biases.
In this section, we introduce a framework supported by iResponse technology and artificial intelligence for automatic management of the pandemic. It allows monitoring and application of pandemic-related policies, planning and provision of resources, data-based planning and decision-making, and coordination in various fields. We describe an iResponse framework in Figure 3, which consists of five modules: chain monitoring and interruption (MBC), cure development and treatment (CDT), resource planner (RP), data analysis, and decision-making (DADM), and Data Storage and Management (DSM).
Provided here is the technical system architecture of the iResponse framework in Figure 4. Various data sources are used to collect data, such as IoT-based sensors, social media, electronic medical records (EHR), hospital occupancy data, Wi-Fi, GPS, travel itineraries, test laboratories, smart devices, etc. In addition, the data is processed. For example, if necessary, missing values and outliers can be identified and predicted. For text-related data, we removed all symbols, special characters, etc. to create a pure text corpus that can be further analyzed. Data is stored in cloud-based data centers as well as internal data centers. These data centers are heterogeneous, which means they can store various data, such as structured, unstructured, multimedia, space, and EHR. There are several machine learning and deep learning algorithms available, which are trained on the data provided by the data center. In addition, these well-trained models can use real-time data to perform specific tasks. Use the data visualization function to visualize the obtained results (data center data and real-time data) statically and interactively.
Conclusion
In this article, we discussed many important findings of the epidemic. The evolution of epidemics varies greatly between countries, and these trends depend on the specific characteristics of the virus itself and the differences in the response measures of countries around the world. As of April 2021, in terms of the number of cases, the five most affected countries are the United States, Brazil, India, Russia, and South Africa. In terms of the death toll, the five most affected countries are the United States, Brazil, Mexico, the United Kingdom, and India. According to the Eurasian Group, which has developed a method to evaluate national responses in three key areas (health management, policy responses, and financial policy responses), the five countries with the best global response to COVID-19 are Taiwan, Singapore, and South Korea. , New Zealand and Australia. Lessons learned from good practices in different countries include the use of physical isolation (also known as social isolation), isolation, lockdown, curfew, and provision of disinfectants to quickly control the virus. Implement containment measures through contact tracing and other methods; increase testability; treatments are not necessarily identified and discovered in pharmaceutical laboratories, but through interaction with the wider community.
Provided here are case studies to elaborate on the different functionalities of the iResponse framework and how the framework can be implemented. These include a sentiment analysis case study, a case study on the recognition of human activities, and studies using deep learning and other data-driven methods to show how to develop sustainability-related optimal strategies for pandemic management using seven real-world datasets. The datasets we used are real-world open datasets including a human activity recognition (HAR) dataset, the World Bank COVID-19 dataset, Google COVID-19 mobility report dataset, credit card transactional data by the U.S.
