About Me
My research seeks to relate how humans and machines learn through the integration of artificial intelligence methods into cognitive models of human learning and decision making.
One application of this research is in better understanding the fundamental properties of human learning, and how they make decisions in a variety of settings. The decision making settings I have worked on in the past include natural language in phishing email categorization, network traffic in cyberdefense, and abstract visual information in utility learning tasks. In these contexts, artificial intelligence methods allow for analysis of complex stimuli like visual information and natural language, which traditional approached in cognitive modeling have difficulty with.
The other main application of this research is in furthering the development of artificial intelligence methods and the interfaces that humans use to interact with these methods. The inspiration of this application of research is that AI methods can benefit from being more human-inspired, if they are going to interact with humans or be used by them, especially by non-experts. Projects that have investigated this area of my research have involved the prompting of large language models with information obtained by classical cognitive models, and encouraging more human-like behavior from reinforcement learning models by imposing a constraint on the complexity of their behavior.
The future areas of this research will be extended into additonal and more complex domains, to understand the progression of artificial intelligence methods, how humans interact with them, and what these methods can tell us about the inner workings of the human mind.