Key Artificial Intelligence Projects to Fight Against COVID-19
Dr. Amit Ray
Compassionate AI Lab
Prevention and early healing are the primary requirements for the present COVID crisis. In our Compassionate AI Lab, we broadly classified our fight against COVID 19 Artificial intelligence (AI) based research projects into six groups. They are AI for COVID vaccine development, AI for COVID drug discovery, AI for COVID diagnosis, AI for COVID testing, AI for COVID growth rate forecasting, and AI for social robots.
Researchers across the world are in search of urgent drugs and vaccines that can save millions of lives of infected people and perhaps prevent infections for the future generations. The exacerbated time, cost and high failure rate of traditional path of drug discovery and vaccine development has prompted the need for efficient use of machine learning techniques. In these projects, we are trying to solve one the most complex problems of humanity ever encountered
1. AI for COVID 19: An Overview
The massive outbreak of the COVID-19 has prompted various scientists, researchers, laboratories, and organizations around the world to conduct large-scale research to help develop vaccines and other treatment strategies. Biology and medicine of coronavirus are data rich, complex, and often ill understood. Problems of this nature may be particularly well suited to deep learning techniques.
Deep learning algorithms are general non-linear models, which are able to learn features directly from data, making them an excellent choice for various COVID applications. There are many AI algorithms available. A brief overview of these deep learning algorithms and deep reinforcement learning are provided here. We examine applications of machine learning to a variety of COVID problems— fundamental biological processes, vaccine development, drug development, social robots, and treatment of patients.
The availability of large COVID biological datasets has promising potential to solve many aspects of COVID problems. The data of SARS-CoV-2 nucleotide sequences from the NCBI, and GISAID databases, are very useful. However, the power of deep learning has not yet been fully utilized. It will definitely help to solve many COVID problems.
Acceleration of drug discovery, vaccine development, chest X-ray or CT image analysis, diagnosis, testing and early infection detection are key areas of AI applications. Among these research works, some of them like COVID growth rate forecasting and COVID CT-image analysis are very easy, and some of them are more complex and need considerable research effort. Due to the recent explosion in biological ’omics’ data, deep learning has also found its application within the biology field; it can be used to COVID in problems such as protein-protein interaction modeling.
2. Key AI Algorithms for COVID Problems
Traditional machine learning methods include logistic regression, decision tree, random forest, K-nearest neighbor, Adaboost, K-means clustering, density clustering, hidden Markov models, support vector machine, XGBoost, Naive Bayes, etc. are useful for classification, regression, clustering, dimensionality reduction and time series analysis. The deep learning algorithms include convolutional neural networks, deep belief networks, restricted Boltzmann machines, recurrent neural networks and reinforcement learning.
Advanced machine learning algorithms such as Generative Adversarial Networks (GANs), Extreme Learning Machine (ELM), Transfer learning, Autoencoder, Polynomial neural networks, Long /Short Term Memory (LSTM), and Recurrent Neural Networks (RNNs) can integrate and analyze large-scale data related to COVID-19 patients to facilitate a deeper understanding of viral spread pattern, improve the speed and accuracy of diagnosis, develop fresh, effective therapeutic approaches, and even identify individuals who, depending on their genetic and physiological features, are most susceptible to the disease.
Natural language processing (NLP) can be used to automatically find meaningful information from unstructured text and speeches. It can be used to improve the efficiency of COVID analysis, patient care and human–social-robot interactions. It is used to assess potential adverse drug reactions.
3. AI for COVID 19 Vaccine Development
The development of reliable and potent vaccinations is the only viable way of ending the COVID-19 pandemic. Currently, there are mainly three types of COVID-19 vaccine candidates, such as (1) whole virus vaccines, (2) recombinant protein subunit vaccines, and (3) nucleic acid vaccines. We focus to develop AI based online tools that can provide functions such as SARS-CoV-2 genetic variation analysis, epitope prediction, coronavirus homology analysis, and candidate proteome analysis. The presence of a large number of experimental and biological databases containing relevant screened compounds are easily accessible via a public domain. Among them, the most widely available databases are ChEMBL and PubChem. Vaccine candidate analysis, epitope prediction, sequence Homology, proteome analysis, and dose response analysis are our main areas for the AI applications for COVID-19 vaccine development research.
4. AI for COVID 19 Drug Discovery
Development of new drugs is a time-consuming and costly process. However, systems biology and machine learning approaches are continuously enhanced in order to accelerate the path to efficient drug development. The high failure rate of drug development pipelines, often at late stages of clinical testing, has always been a critical issue. Indeed, in order to ensure both the patients’ safety and drug effectiveness, prospective drugs must undergo a several stages of validity checks and sensitivity tests. Machine learning techniques can reduce the drug candidate identification phase substantially. We focused primarily on using Generative Adversarial Network (GAN) model for COVID drug discovery. Generative Adversarial Network is a set of two neural networks, the Generator and the Discriminator. These two networks are trained at the same time.
Recently, DNA sequencing technology adopted machine learning to read out long stretches of DNA fragments from digital electronic signaling data. Long read technologies are important to resolve repetitive regions in the genome and detect complex structural variants. The current short technology can not resolve these issues and it is still unknown the disease risk contribution from repetitive region and structural variation of the genome. Target identification, compound design, drug testing, drug re-purposing, and drug response analysis are our key areas for the AI applications for COVID-19 drug discovery research.
5. AI for COVID 19 Diagnosis
COVID diagnosis broadly cover various laboratory testing, medicine, therapy, prevention and life support. Rapid diagnosis of the COVID-19 can allow governments to take effective response measures to limit the disease’s further spread. AI technology has considerable potential to improve image-based and symptom based medical diagnosis. A majority of the COVID patient exhibited clinical characteristics such as fever, dry cough, fatigue, and sputum production. At the same time, some patients showcased symptoms such as sore throat, headache, myalgia, and breathlessness.
For the diagnosis of coronavirus, Reverse Transcriptase Polymerase Chain Reaction (RT–PCR) test is routinely used. However, this test can take up to two days to complete, serial testing may be required to rule out the possibility of false negative results and there is currently a shortage of RT–PCR test kits, underscoring the urgent need for alternative methods for rapid and accurate diagnosis of patients with COVID-19.
Chest X-ray or CT image analysis is a valuable component of evaluation and diagnosis in symptomatic patients with suspected SARS-CoV-2 infection. AI algorithms may meet this need by integrating chest CT image analysis findings with clinical symptoms, exposure history and laboratory testing in the algorithm.
6. AI for COVID 19 Testing
Developing accurate and reliable tests to diagnose SARS-CoV-2 infection in individuals is essential to curb its rapid transmission. The way to slow down the spread of the pandemic is to increase the number of tests and break the chain of infection. AI models can be developed to predicted with high accuracy whether an individual is likely to have COVID-19 based on their age, sex and a combination of various key symptoms like: loss of smell or taste, severe or persistent cough, fatigue and skipping meals. The distinction between diagnosed and non-diagnosed is important because non-diagnosed individuals are more likely to spread the infection than diagnosed ones.
The currently available COVID-19 tests can be broadly classified into two classes: Molecular tests, Serology tests. Serology tests are blood-based tests that can be used to identify whether people have been exposed to a particular pathogen by looking at their immune response.
7. AI for COVID 19 Growth Rate Prediction
The early predictions might help to prepare against possible threats and consequences. Classic epidemic models are also useful to obtain mathematical models for epidemics. However, many parameters of these models, such as infected rate and basic reproduction number, medical (symptomatic and asymptomatic) parameters, require data-driven approaches to estimate them accurately. Due to a high level of uncertainty and lack of essential data, standard models have shown low accuracy for long-term prediction. Comparatively AI based prediction models have many advantages.
8. AI for COVID 19 Social Robots
Compassionate social robots is one of the key research area of our Compassionate AI Lab. AI based social robots for COVID disinfection, patient care, oxygen & drug support, sample collection, delivery robots are key areas of AI applications. Many countries have used smart devices equipped with AI to detect suspicious persons in public transportation places such as airports and train stations. For example, infrared cameras are used to scan for high temperatures in a crowd, and different AI methods perform efficient analysis to detect whether an individual is wearing a mask in real time. AI and NLP technologies can be used to develop remote video diagnosis systems and chat robot systems and provide COVID-19 disease consultation and preliminary diagnosis to the public.
9. Conclusion and Discussion
Artificial Intelligence can conceivably play an essential role in mitigating the impact of the COVID-19 pandemic. However, at present, AI systems are still in the prefatory stages. The several challenges and limitations hindering the application of AI in COVID-19. I am inspired and encouraged by the speed at which the students, researchers, volunteers, and research organizations are applying machine learning to address COVID-19 issues. I believe in the potential of machine learning to help solve the biggest challenges in our world – and that promise is now coming to fruition as organizations are responding to this crisis with huge supports. It is my hope that in this difficult time we can work together on a global scale to innovate and find new ways and machine learning can contribute a lot in this fight against COVID-19.
Cite This ArticleTo export a reference to this article please select a referencing style below:
|title=Artificial Intelligence to Fight Against COVID-19
|publisher= Inner Light Publishers
|access-date= 04 Mar 2024