Completed
Pseudo-dimension (for real valued classes)
Class Central Classrooms beta
YouTube videos curated by Class Central.
Classroom Contents
Learning Theoretic Foundations of Data-Driven Algorithm Design
Automatically move to the next video in the Classroom when playback concludes
- 1 Intro
- 2 Machine Learning for Algorithm Design Learning algorithms for solving combinatorial problems. E.g.
- 3 Data-driven Algorithm Design Data-driven algo design: use learning & data for algo design. Suited when repeatedly solve instances of the same algo problem
- 4 Data-driven algorithm design: Problem Setup
- 5 Uniform Convergence Uniform convergence for any algo in Als, average performance over samples close to its expected performance. • Imply that Ā that does best over the sample has high expected perfor…
- 6 General Sample Complexity via Dual Classes
- 7 Pseudo-dimension (for real valued classes)
- 8 Pseudo-dimension, Uniform Convergence
- 9 Online Algorithm Selection (via online optimization of piecewise Lipschitz functions)
- 10 Online Regret Guarantees Existing techniques (for finite, linear, or convex case) select
- 11 Summary and Discussion Data-driven algo design can overcome major shortcomings of classic design by adapting the algo to the domain at hand. Different methods work better in different settings Learn …
