Interpretable Explanations of Black Boxes by Meaningful Perturbation - CAP6412 Spring 2021

Interpretable Explanations of Black Boxes by Meaningful Perturbation - CAP6412 Spring 2021

UCF CRCV via YouTube Direct link

Conclusion

22 of 23

22 of 23

Conclusion

Class Central Classrooms beta

YouTube videos curated by Class Central.

Classroom Contents

Interpretable Explanations of Black Boxes by Meaningful Perturbation - CAP6412 Spring 2021

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

  1. 1 Intro
  2. 2 Content
  3. 3 Abstract Image Saliency Methods Summary Attention Map Limited by heuristic properties and architectural constraints
  4. 4 Introduction Current Problems The interpretation for the black box predictor The intuitive visualization method is only heuristic, and the meaning remains unclear.
  5. 5 Contribution Develop principles and methods to explain any black box function By determine mapping attributes - Internal mechanisms is used to implement these attributes
  6. 6 Related Work Gradient-based method -Backpropagates the gradient for a class label to the image layer Other methods: DeConvNet, Guided Backprop
  7. 7 Related Work - CAM
  8. 8 Related Work Comparison
  9. 9 Comparison with other saliency methods
  10. 10 Principle Black bax is a mapping function
  11. 11 Explanations as meta-predictors Rules are used to explain a robin classifier
  12. 12 Advantages of Explanations as Meta-predictors The faithfulness of images can be measured as prediction accuracy To find the explanations automatically
  13. 13 Local Explanations
  14. 14 Saliency Deleting parts of image x, as the perturbations for the whole image X
  15. 15 A Meaningful Image Perturbation 11
  16. 16 Deletion and Preservation
  17. 17 Artifacts Reduction
  18. 18 Experiment-Interpretability
  19. 19 Experiment Testing hypotheses: animal part saliency
  20. 20 Experiment-Adversarial defense
  21. 21 Experiment localization and pointing
  22. 22 Conclusion
  23. 23 Questions?

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.