Optimizing for Interpretability in Deep Neural Networks - Mike Wu

Optimizing for Interpretability in Deep Neural Networks - Mike Wu

Stanford MedAI via YouTube Direct link

Intro

1 of 23

1 of 23

Intro

Class Central Classrooms beta

YouTube videos curated by Class Central.

Classroom Contents

Optimizing for Interpretability in Deep Neural Networks - Mike Wu

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

  1. 1 Intro
  2. 2 The challenge of interpretability
  3. 3 Lots of different definitions and ideas
  4. 4 Asking the model questions
  5. 5 A conversation with the model
  6. 6 A case for human simulation
  7. 7 Simulatable?
  8. 8 Post-Hoc Analysis
  9. 9 Interpretability as a regularizer
  10. 10 Average Path Length
  11. 11 Problem Setup
  12. 12 Tree Regularization (Overview)
  13. 13 Toy Example for Intuition
  14. 14 Humans are context dependent
  15. 15 Regional Tree Regularization
  16. 16 Example: Three Kinds of Interpretability
  17. 17 MIMIC III Dataset
  18. 18 Evaluation Metrics
  19. 19 Results on MIMIC III
  20. 20 A second application: treatment for HIV
  21. 21 Distilled Decision Tree
  22. 22 Caveats and Gotchas
  23. 23 Regularizing for Interpretability

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.