DeepAI AI Chat
Log In Sign Up

Fast Geometric Projections for Local Robustness Certification

02/12/2020
by   Aymeric Fromherz, et al.
0

Local robustness ensures that a model classifies all inputs within an ϵ-ball consistently, which precludes various forms of adversarial inputs. In this paper, we present a fast procedure for checking local robustness in feed-forward neural networks with piecewise linear activation functions. The key insight is that such networks partition the input space into a polyhedral complex such that the network is linear inside each polyhedral region; hence, a systematic search for decision boundaries within the regions around a given input is sufficient for assessing robustness. Crucially, we show how these regions can be analyzed using geometric projections instead of expensive constraint solving, thus admitting an efficient, highly-parallel GPU implementation at the price of incompleteness, which can be addressed by falling back on prior approaches. Empirically, we find that incompleteness is not often an issue, and that our method performs one to two orders of magnitude faster than existing robustness-certification techniques based on constraint solving.

READ FULL TEXT

page 1

page 2

page 3

page 4

07/23/2020

Hierarchical Verification for Adversarial Robustness

We introduce a new framework for the exact point-wise ℓ_p robustness ver...
02/28/2022

An Analytical Approach to Compute the Exact Preimage of Feed-Forward Neural Networks

Neural networks are a convenient way to automatically fit functions that...
02/23/2021

Non-Singular Adversarial Robustness of Neural Networks

Adversarial robustness has become an emerging challenge for neural netwo...
01/04/2020

Empirical Studies on the Properties of Linear Regions in Deep Neural Networks

A deep neural network (DNN) with piecewise linear activations can partit...
10/29/2019

Learning Without Loss

We explore a new approach for training neural networks where all loss fu...
05/17/2021

How to Explain Neural Networks: A perspective of data space division

Interpretability of intelligent algorithms represented by deep learning ...
03/20/2019

Provable Certificates for Adversarial Examples: Fitting a Ball in the Union of Polytopes

We propose a novel method for computing exact pointwise robustness of de...