Course Identification
Seminar on Machine Learning Theory
Lecturers and Teaching Assistants
Prof. Ohad Shamir
Course Schedule and Location
Sunday, 09:15 - 11:00, Ziskind, Rm 1
03/11/2024
Field of Study, Course Type and Credit Points
Mathematics and Computer Science: Seminar; Elective; Regular; 2.00 points
Attendance and participation
Required in at least 80% of the lectures
Estimated Weekly Independent Workload (in hours)
Syllabus
In this seminar, students will read, present and discuss recent research papers in theoretical machine learning, in particular deep learning and related topics. The focus will be on rigorous results, pertaining to questions such as why predictors can be successfully trained with simple gradient based methods; why large-scale models (e.g. modern neural networks) are able to generalize well; and how does the predictor's architecture affects the types of functions it can express. Students will be able to choose a paper from a predefined list, or present a paper of their own choice (in coordination with the instructor).
Learning Outcomes
Upon succesful completion of this course:
The students will gain familiarity with the current state-of-the-art in machine learning theory.