April 18, 2023
Dr. Na Du, an assistant professor in the Department of Informatics and Networked Systems, is trying to develop a better understanding of how humans interact with autonomous machines. “My research would like to bridge the gap between increasingly complex autonomous technologies and restricted human capacities in information processing, cognition, and decision-making,” Du said. Through closely examining human behaviors within autonomous systems, Du hopes to develop personalized technologies that adapt to users’ real-time cognitive and physical states.
Informed by her background in industrial and engineering psychology, Du combines the theoretical insights of those fields with data-driven methods “to understand and model human behaviors in intelligent systems.” Du said, “Some people do psychology, some do data-driven methods. I would like to combine both the data-driven methods and what we call cognitive architecture to better understand and model human behaviors in intelligent systems.”
Reflecting on her experience as a researcher, Du said "I realized how vulnerable humans are when they interact with complex systems. There are lots of fallacies, illusions, and biases in human perception, information processing, and decision-making, which makes their interaction with technology worse than they expected. This is not the fault of the human. The technologies should be designed to fit humans.”
Du remains busy, with several projects currently underway. Each project concerns how to design better technologies for users by focusing on how they process information while using different tools. Du’s attention has most recently turned to autonomous vehicles. The aim of that project, Du said, is “to model the driver’s situational awareness within an automated vehicle, which means how aware you are of the surrounding environment. Here, we would like to use different wearable sensors, collecting the driver’s eye movement, heart rate and skin conductance for human-centered computing. And then we would like to develop an alert system that could support the driver to take control of the vehicle when the automated vehicle is not able to drive by itself.” Another project seeks to “design a ‘serious game’ to help users understand privacy policy, rather than blindly accept all of the privacy policies.” Finally, Du is conducting research that could be used to help design more accessible public spaces for people with disabilities. She said, “We tried to develop an infrastructure for navigating outdoor environments, an infrastructure for people who have disabilities. And here, infrastructure can be understood as the machine learning algorithms or simulations that generate accessible routes and navigation for a pedestrian. It could be a mobile application or some design requirements or design suggestions for the cyberinfrastructure in indoor or outdoor environments.”
Students have the opportunity to immerse themselves in studying these approaches in Du’s graduate courses. In her class titled “Human Information Processing, Du said she “tried to employ existing theoretical knowledge to real-world applications. I used special topics to help students understand how to apply what they have learned towards real-world situations.” Concerning the skills that students will develop, Du said “The skillset that they practice in the projects could be about data analytics, experimental design like A/B testing, user experience design, or human-centered computing. And for the topic side, it’s more about human-computer interaction and how the human interacts with intelligent systems. If they want to practice programming skills in data analytics, and they’re interested in programming the test bad, and they’re interested in those intelligent systems and the human interaction with Intelligent systems, there are some research opportunities here.”
--Daniel Beresheim