From Learning to Discovery: Lessons from a Semester of AI-Assisted Optimization Seminar

Can AI enable students to contribute to research before completing the traditional path to expertise? And can working on research open problems itself become an effective mode of learning? In a semester-long seminar, teams of bachelor’s and master’s students used AI tools to investigate nine open problems in optimization. Drawing on both successful and unresolved cases, this talk explores how AI can support just-in-time learning, accelerate mathematical exploration, and enable earlier participation in research.