Explores the formulation and complexity of Support Vector Machines, including primal and dual forms, geometric interpretation, and algorithmic implications.
Covers transformer architecture and subquadratic attention mechanisms, focusing on efficient approximations and their applications in machine learning.