Machine Learning Tools for Computational Cognitive Modeling
Anne Collins, Professor
Psychology
Applications for Fall 2025 are closed for this project.
Computational cognitive models are mathematical tools that help us understand the underlying processes of human cognition. Researchers use these models to test different hypotheses about how the mind works. Traditionally, selecting the best model to explain human behavior relies on standard statistical techniques like maximum likelihood estimation. However, these methods often fall short when dealing with models that have intractable likelihoods.
Recent advances in machine learning (ML) offer new avenues to address these challenges. This project will introduce students to a range of cognitive models commonly used to study human learning behavior.
Role: Students will work closely with a mentor to apply both traditional statistical methods and modern ML approaches to compare and evaluate these models using behavioral data.
In addition to hands-on modeling work, students will explore relevant literature on human cognition and the integration of ML into cognitive science. They will also participate in broader lab activities, including regular meetings with the research team and the principal investigator.
Qualifications: Applicants must have a quantitative background, including statistics, mathematics, computer science and/or machine learning, as well as an interest in cognitive science. Experience with the following is a plus: training recurrent neural networks. We also have a preference for students who expect to participate in the lab for more than a single academic year. Students must be curious, self-driven, independent, and ready to take on new challenges.
Day-to-day supervisor for this project: Ti-Fen Pan, Graduate Student
Hours: 9-11 hrs
Related website: https://ccn.berkeley.edu/
Social Sciences Education, Cognition & Psychology