Coursera Spring Sale
40% Off Coursera Plus Annual!
Grab it
Attend a machine learning lecture exploring fast agnostic learners for geometric concept classes in the plane. Discover recent research results on the time complexity of agnostic learning, which has been far less explored compared to its well-understood sample complexity. Learn about proper agnostic learners for triangles, convex k-gons (for small k), and convex sets in the square, including optimal sample complexity approaches that improve running times for triangles, 4-gons, and 5-gons. Explore an agnostic learner for convex sets under uniform distribution that achieves faster computation with only minor costs in sample complexity. Examine the connections between agnostic learning and tolerant property testing through research conducted jointly with Ludmila Glinskih and Sofya Raskhodnikova. The presentation is delivered by Talya Eden, Assistant Professor at Bar Ilan University's computer science department, who specializes in sublinear-time and randomized graph algorithms for huge datasets, with particular focus on theoretical and applied graph parameter estimation in sublinear-time and related areas including streaming models, massively parallel computation, and learning-augmented algorithms.
