18/07/2021

Provable Lipschitz Certification for Generative Models

Matt Jordan, Alex Dimakis

Keywords: Social Aspects of Machine Learning, Privacy, Anonymity, and Security

Abstract Paper Similar Papers

Abstract: We present a scalable technique for upper bounding the Lipschitz constant of generative models. We relate this quantity to the maximal norm over the set of attainable vector-Jacobian products of a given generative model. We approximate this set by layerwise convex approximations using zonotopes. Our approach generalizes and improves upon prior work using zonotope transformers and we extend to Lipschitz estimation of neural networks with large output dimension. This provides efficient and tight bounds on small networks and can scale to generative models on VAE and DCGAN architectures.

 0
 0
 0
 0
  Share    
This is an embedded video. Talk and the respective paper are published at ICML 2021 virtual conference. If you are one of the authors of the paper and want to manage your upload, see the question "My papertalk has been externally embedded..." in the FAQ section.

Comments

Post Comment
no comments yet
code of conduct: tbd Characters remaining: 140

Similar Papers

 0
 0
 0
 0
 5:00 
 0
 0
 0
 0
 5:34 
 0
 0
 0
 0
 15:38 
 0
 0
 0
 0
 3:17 
 0
 0
 0
 0
 15:02 
 0
 0
 0
 0
 13:12 
 0
 0
 0
 0
 10:15 
 0
 0
 0
 0
 5:14 
 0
 0
 0
 0
 18:48 
 0
 0
 0
 0
 3:04 
 0
 0
 0
 0
 3:26 
 0
 0
 0
 0
 8:23 
 0
 0
 0
 0
 5:25 
 0
 0
 0
 0
 13:29 
 0
 0
 0
 0
 14:07 
 1
 1
 0
 0
 5:18 
 0
 0
 0
 0
 3:02 
 0
 0
 0
 0
 18:54 
 0
 0
 0
 0
 14:59 
 0
 0
 0
 0
 15:27 
 0
 0
 0
 0
 15:40 
 0
 0
 0
 0
 13:48 
 0
 0
 0
 0
 14:05 
 0
 0
 0
 0
 3:28 
 0
 0
 0
 0
 3:25 
 0
 0
 0
 0
 12:19 
 0
 0
 0
 0
 4:44 
 0
 0
 0
 0
 18:46 
 0
 0
 0
 0
 4:48 
 0
 0
 0
 0
 3:21 
 0
 0
 0
 0
 5:20 
 0
 0
 0
 0
 8:40 
 0
 0
 0
 0
 15:08 
 0
 0
 0
 0
 14:50 
 0
 0
 0
 0
 9:07 
 0
 0
 0
 0
 15:10 
 0
 0
 0
 0
 3:04 
 0
 0
 0
 0
 14:37 
 0
 0
 0
 0
 3:27 
 0
 1
 1
 0
 5:13