Regularization for Optimal Transport and Dynamic Time Warping Distances - Marco Cuturi

Regularization for Optimal Transport and Dynamic Time Warping Distances - Marco Cuturi

Alan Turing Institute via YouTube Direct link

OT as a Loss: Differentiability?

16 of 29

16 of 29

OT as a Loss: Differentiability?

Class Central Classrooms beta

YouTube videos curated by Class Central.

Classroom Contents

Regularization for Optimal Transport and Dynamic Time Warping Distances - Marco Cuturi

Automatically move to the next video in the Classroom when playback concludes

  1. 1 Intro
  2. 2 Dynamic Time Warping
  3. 3 Pairwise Distance Matrix
  4. 4 Alignment Path
  5. 5 Path Cost
  6. 6 Min Cost Alignment Matrix?
  7. 7 Best Alignment Matrix
  8. 8 Best Path: Bellman Recursion
  9. 9 Optimal Path
  10. 10 OT for Discrete Measures
  11. 11 Wasserstein on Discrete Measures
  12. 12 Dual Kantorovich Problem
  13. 13 Solving the OT Problem
  14. 14 In Summary
  15. 15 DTW as a Loss: Differentiability?
  16. 16 OT as a Loss: Differentiability?
  17. 17 Any way to fix this?
  18. 18 Example softmin of quadratic functions
  19. 19 Recursive Computation (Backward)
  20. 20 Computation Graph: Forward
  21. 21 Backward Recurrence
  22. 22 Generating Function for OT
  23. 23 Fast & Scalable Algorithm
  24. 24 Sinkhorn as a Dual Algorithm
  25. 25 Block Coordinate Ascent, a.k.a Sinkhorn
  26. 26 Differentiability of W
  27. 27 Algorithmic Formulation
  28. 28 Sinkhorn: A Programmer View
  29. 29 Interpolation Between 2 Time Series

Never Stop Learning.

Get personalized course recommendations, track subjects and courses with reminders, and more.

Someone learning on their laptop while sitting on the floor.