Machine learning applied to COVID-19

Alderisi, Serena (2020). Machine learning applied to COVID-19. Thesis (Master thesis), E.T.S. de Ingenieros Informáticos (UPM).


Title: Machine learning applied to COVID-19
  • Alderisi, Serena
  • Larrañaga Múgica, Pedro
  • Bielza Lozoya, María Concepción
Item Type: Thesis (Master thesis)
Masters title: Ciencia de Datos
Date: October 2020
Faculty: E.T.S. de Ingenieros Informáticos (UPM)
Department: Inteligencia Artificial
Creative Commons Licenses: Recognition - No derivative works - Non commercial

Full text

PDF - Requires a PDF viewer, such as GSview, Xpdf or Adobe Acrobat Reader
Download (2MB) | Preview


This work is focused on the impact of machine learning, on the COVID-19 pandemic. Machine learning has proven to be invaluable in predicting risks in many spheres and since the spread of the virus started, its application is helping us against the viral pandemic. Like never before, people all around the world are collecting and sharing what they learn about the virus. Hundreds of research teams are combining their efforts to collect data and develop solutions every day. Starting from this, the main goals of this work are: to shine a light on their work; going deep into how the application of machine learning techniques on different fields affected by the pandemic is helping us in the fight against the coronavirus; to identify strengths and weaknesses of machine learning techniques and the challenges for further progress in medical machine learning systems. This final master thesis report addresses recent studies that apply machine learning on multiple angles: screening and diagnosis, contact tracing, drugs/vaccines development and prediction and forecasting for COVID-19. Early and timely treatment computer-aided diagnosis, highlighted the importance of machine learning algorithms that proved to be urgently needed. Its application could largely reduce the efforts of clinicians and accelerate the screening and diagnosis process, as well as, reduce the human intervention in medical practice. Analyzing X-rays images with deep learning algorithms makes possible to achieve amazing results: in some cases models performance surpass the ability of a senior radiologists to early distinguish different types of pneumonia from COVID-19 pneumonia. In addition, deep neural networks are not only a valuable support to the activity of radiologists but they also represent a double check for false RT-PCR responses. Using mobile apps for contact tracing helps to make people more aware about the consequences of their contacts and their movements. This consciousness will allow to contain the spread of the virus and individual protection. Having situational awareness by governments would benefit from increased access to previously unavailable population estimates and mobility information to enable different sectors to better understand COVID-19 trends and geographic distribution. This could be a big improvement, since the lack of available data is a big limitation for the implementation of machine learning solutions in the real world. Machine learning application on laboratory research is allowing to speed up a vaccine development against SARS-CoV-2 virus and to discover which among the already existing drugs have potential anti-coronavirus activities. Last but not least, knowing earlier who, among the cases, will develop severe symptoms, allows to manage better medical resources and treatments and mostly to avoid intensive care unit stress. Furthermore, knowing the possible development of the contagion dynamics lets governments plan more efficient and less-invasive containment strategies. The outcome of this work can provide deep insights into this disease, re-evaluate existing medical diagnosis systems for this virus, and recommend solutions for developing reliable, faster and more efficient medical systems, with the aid of machine learning application.

More information

Item ID: 65544
DC Identifier:
OAI Identifier:
Deposited by: Biblioteca Facultad de Informatica
Deposited on: 24 Nov 2020 15:42
Last Modified: 20 May 2022 17:05
  • Logo InvestigaM (UPM)
  • Logo GEOUP4
  • Logo Open Access
  • Open Access
  • Logo Sherpa/Romeo
    Check whether the anglo-saxon journal in which you have published an article allows you to also publish it under open access.
  • Logo Dulcinea
    Check whether the spanish journal in which you have published an article allows you to also publish it under open access.
  • Logo de Recolecta
  • Logo del Observatorio I+D+i UPM
  • Logo de OpenCourseWare UPM