Embedded Files
We are recruiting graduate students
We are recruiting graduate students
You may apply through School of Digital Humanities & Computational Social Sciences (디지털인문사회과학부), Department of Brain & Cognitive Sciences (뇌인지과학과), or Graduate School of Data Science (데이터사이언스대학원) at KAIST. Please contact the PI to check availability in each department.
Research Directions
Research Directions
Interaction between psychology and artificial intelligence (AI) is intriguing because they can learn from each other. AI models specialized in predicting complex data provide innovative ways to explain human behaviors in complex real world. Understanding the human mind helps AI implement efficient functions inspired by human brain processes, act like humans, and collaborate with humans. We explore the intersection of psychology and AI through the following questions.
AI in Psychology: How can we use artificial intelligence models to explain human behaviors?
Psychology in AI: Can we improve AI by incorporating human cognitive functions?
To address these questions, we employ a variety of methods including Computational Cognitive Modeling, Deep Learning, and (Inverse) Reinforcement Learning.
AI in Psychology
AI in Psychology
Human Performance
AI Performance
A recent breakthrough in psychology is the use of AI for data analysis and computational modeling. Still, most studies use AI merely as an analysis tool, rather than a building block of computational models that explain human cognition and behavior. Using AI as a cognitive model is challenging because of inherent differences in behaviors between AI and humans.
We aim to explain individual differences in human behaviors and characteristics by training and interpreting AI agents that imitate complex human behaviors in video games and virtual reality.
Psychology in AI
Psychology in AI
Human cognitive functions are remarkably efficient. For example, attention mechanisms quickly filter the vast amount of information from the world and efficiently allocates attention to relevant stimuli. Imitating human attention mechanisms has enhanced the efficiency and explainability of AI algorithms, suggesting that developing AI based on theories of human cognition could be an innovative approach to overcoming the limitations of data-driven AI algorithms. However, AI researchers often overlook the details of human cognition that psychologists have studied.
We aim to develop explainable AI based on theories from cognitive psychology and data from human experiments.
Computational Cognitive Modeling
Computational Cognitive Modeling
What mathematical functions in our brain determine our behaviors? How can we quantify cognitive traits and characteristics? Computational cognitive modeling is a technique to develop computational/mathematical models that explain cognitive processes underlying human behaviors and brain activity. This approach serves as the foundation of psychological AI in our lab, as well as digital healthcare and cognitive training in the real world.
Questions?
Questions?
Contact sangholee@kaist.ac.kr
to get more information.
Page updated
Google Sites
Report abuse