A New Perspective on Adversarial Perturbations

A New Perspective on Adversarial Perturbations

Simons Institute via YouTube Direct link

The Role of Robust Training

12 of 18

12 of 18

The Role of Robust Training

Class Central Classrooms beta

YouTube playlists curated by Class Central.

Classroom Contents

A New Perspective on Adversarial Perturbations

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

  1. 1 Why do we love deep learning?
  2. 2 Key Problem: Adversarial Perturbations
  3. 3 ML via Adversarial Robustness Lens
  4. 4 Human Perspective
  5. 5 ML Perspective
  6. 6 A Simple Experiment
  7. 7 The Robust Features Model
  8. 8 The Simple Experiment: A Second Look
  9. 9 Human vs ML Model Priors
  10. 10 New capability: Robustification
  11. 11 A Natural Consequence: Transferability
  12. 12 The Role of Robust Training
  13. 13 New Take on Randomized Smoothing
  14. 14 Robustness and Data Efficiency
  15. 15 A Simple Theoretical Setting: Max Likelihood Gaussian Classification
  16. 16 Robustness + Perception Alignment
  17. 17 Robustness + CV Applications
  18. 18 Adversarial examples arise from non-robust features in the data

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.