6.S986: Uncertainty Quantification with AI, Spring 2026OverviewThis is a topics course on modern research work related to Uncertainty Quantification with AI. The aim of this course is to discuss contemporary research in this topic for graduate students working in related areas. The course will focus on how core topics in calibration, distributional prediction, and conformal prediction relate to modern AI systems, such as LLMs and agentic AI. FormatThe first 6 lectures will be given by the instructor, and there will be two homeworks assigned during this time. The remaining sessions will be presentations of recent research papers given by students. All students are required to present one time during the semester. Presenters may select a paper to present from a recommended bank or propose a different paper. There are no exams in this course. Time and location
|