Novel AI Can Predict COVID-19 Vaccine Attitude Using Integrated System of Machine Learning, Math of Human Judgment

A new artificial intelligence (AI) tool can predict a person's willingness to receive a COVID-19 vaccination.

New AI Predicts Vaccine Attitude

Such findings serve as a framework for novel technology that could be widely applied in mental health predictions and lead to more effective public health campaigns.

The system makes use of a small dataset from personal judgments and demographics. The dataset comprised 15 judgment variables and some demographic variables.

The model was created by a team led by researchers from the University of Cincinnati and Northwestern University. The team made a predictive model with machine learning and a system covering math equations that describe lawful aversion judgment and reward patterns.

Nicole Vike, the study's lead author and a senior research associate from the College of Engineering and Applied Science at UC, explains that they used few variables and minimal computation resources to make such predictions.

Vike explains that it is unlikely that COVID-19 will be the last pandemic in the coming decades. A novel, predictive AI for public health could be valuable in preparing hospitals to predict the rates of vaccination and the consequent rates of infection.

Findings were noted in the "Predicting COVID-19 Vaccination Uptake Using a Small and Interpretable Set of Judgment and Demographic Variables: Cross-Sectional Cognitive Science Study" study.

ALSO READ: AI Model Assesses Lung Cancer Risk of Non-Smokers, Never-Smokers by Looking at Just One X-Ray Scan

Using Little Data To Make Accurate Predictions

In 2021, during the COVID-10 pandemic, the scientists surveyed 3,476 adults from all over the US. Back then, the first vaccines were available for over a year.

Survey respondents provided details, such as their area of residence, highest educational attainment, income, internet access, and ethnicity. The demographics of the respondents mirrored those of the US based on figures from the US Census Bureau.

The respondents were also asked if they received any of the available vaccines. Roughly 73% of the participants reported being vaccinated.

They were also asked if they routinely were obliged to follow the recommendations, such as social distancing and wearing a mask, to prevent further spread. They were also asked to assess the degree to which they liked or disliked a set of 48 randomly sequenced pictures. They had to do so on a seven-point scale. Such images were taken from the International Affective Picture Set, a massive set of emotionally evocative color photographs. These photos come in six categories: disasters, sports, aggressive animals, cute animals, food, and nature.

According to Vike, this exercise aimed to quantify the mathematical features of people's judgments as they observe mildly emotional stimuli. For this task, the measures included concepts such as aversion to loss and risk, which behavioral economists are familiar with.

This is the willingness towards risk avoidance. An example of this would be getting insurance.

Hans Breiter, a co-author of the study and computer science professor at UC, explains that the framework for judging what is aversive or rewarding is crucial to medical decision-making. A 2017 seminal paper argued that the mind has a standard model. Such a model can be supported with a small variable dataset from mathematical psychology for medical behavior predictions. This collaborative team's work has offered the needed support and claims that the mind has various equations that can be likened to what is used in particle physics.

Moving forward, the researchers used three machine learning approaches to see how well the participants' demographics, attitudes, and judgment toward precautions for COVID-19 predicted whether they would opt to get vaccinated.

The study shows that AI can accurately predict human attitudes with minimal data and reliance on time-consuming and costly assessments. The study argues that "fewer but more interpretable" variables can be used.

Aggelos Katsaggelos, a co-author of the study and an endowed professor of computer science and electrical engineering at Northwestern University, says that this research is anti-big data. He adds that it can work extremely simply without the need for super-computation and costly expenses. In fact, it can be applied by anyone with a smartphone.

The researchers regard it as computational cognition AI. Katsaggelos says that more applications involving judgment alterations are likely to be seen in the future.

RELATED ARTICLE: Artificial Intelligence (AI) Models With LLM-Based Agents Always Choose War Over Peace [Study]

Check out more news and information on Artificial Intelligence in Science Times.