CAREER Award: Teaching machines the art of human
Children attend preschool and kindergarten to learn the skills they need to thrive. They practice sharing and making friends. They learn to use words and numbers. They gain confidence through movement and self-control. Teachers have a range of techniques to help children develop these skills.
But how should they help struggling students? There is no shortage of opinions from educators, parents and policy makers on effective ways to teach children, based on observations of what happens in classrooms and children’s behavior inside and outside. outside the school. However, these observations may be missing something important – factors that are unseen, unrecognized, or unreported, but have a powerful influence on a student’s performance.
Jundong Liassistant professor of electrical and computer engineering, computer science, and data science at the University of Virginia, conducts research that could help teachers and administrators more accurately determine which learning methods work best for their students. younger students.
Li won a prestigious National Science Foundation CAREER Award for better understanding cause and effect in human decision-making in the age of big data. Li will use his $600,000 prize over five years to develop a suite of sophisticated algorithms and mathematical models, informed by human experience and intuition, to find cause and effect relationships in massive amounts of data. His work has the potential for broad applications in public health and medicine in addition to education.
The CAREER program, one of NSF’s most prestigious awards for early career faculty, recognizes the recipient’s potential for leadership in research and education. Li’s award recognizes his expertise in data mining, machine learning and artificial intelligence, which are part of a research strength area for the Charles L. Brown Department of Electrical and Computer Engineering within the School of Engineering and Applied Science at AVU.
“The fundamental problem here is that machine learning and data mining alone are often insufficient to make decisions for humans,” Li said. machine learning can find correlations and then use those correlations to make inferences and predict outcomes.”
Because machines cannot truly understand human needs, expectations, and behaviors, their predictions and recommendations may be based on false correlations.
“We all know that correlation doesn’t necessarily imply causation,” Li said. “To make a decision, we usually need to have a better understanding of cause and result. We want to find causal relationships between the variables at play.” This means creating what Li calls a causal inference model, which quantifies the strength of cause-effect relationships between different variables and uses the strongest to make a decision.
Research to improve the reasoning of machine learning algorithms and models these days is largely data-driven, Li said. to give the model the benefit of human wisdom when processing data and interpreting decision-making scenarios.”
Li has had preliminary success in his proposed approach, working in public health supported by a RAPID grant from UVA Global Institute for Infectious Diseases. Li collaborated with Daniel Mietchen, formerly at UVA’s School of Data Science and now a researcher at the Fraunhofer Institute for Biomedical Engineering in Germany, to assess the impact of COVID-19-related policies on outbreaks. Three members of Li’s research group participated in the study.
The team’s model shows how COVID-19 policies such as social distancing have affected county-level outbreaks, taking into account people’s alertness to the virus over time.
A county government may issue policies to enforce social distancing early in the pandemic, but if county residents tend to be more alert to COVID-19, they would likely have a lower likelihood of infection. In this case, vigilance is a confounding variable, influencing both the “treatment”, or the social distancing policy, and the outcome, or the number of individuals who fall ill.
Publicly available information online provided an important resource. For example, the team used the popularity of Google searches for COVID-19 at different times as a measure of resident alertness. Using this indicator and others, the team developed a framework that captures information from different time periods and manages information across counties to estimate how various policies have affected COVID-19 outbreaks. The framework shows the cause and effect of policies at different degrees of specificity, from a category of policies with a certain objective, to a single policy.
The team members presented the results of their study in a research paper, Assessing the Causal Impact of COVID-19 Related Policies on Epidemic Dynamics: A Case Study from the United Statespublished in the proceedings of the Association for Computing Machinery Web Conference 2022 in April.
“Our web conference paper captures the dynamics of outbreaks more accurately than statistical methods alone,” Li said. “In addition, our policy assessment is more consistent with existing epidemiological studies on COVID-19. This suggests that public health officials can use our framework when randomized controlled trials, the gold standard of cause-effect estimation, are not feasible.
Li’s next step involves collaborating with people who know the application areas, such as doctors, public health officials, and learning and development experts. It will also identify publicly available data in health and education that it can leverage to test and further develop its decision-making framework.
Ultimately, Li plans to develop sophisticated algorithms that will identify cause and effect, so doctors can use them to personalize treatments based on patient information, and decision makers can connect the algorithms to their own. data systems to propose policies that improve their constituents’ health, economic well-being and quality of life.
Warning: AAAS and EurekAlert! are not responsible for the accuracy of press releases posted on EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.