Towards Global Neural Network Abstractions with Locally-Exact Reconstruction

10/21/2022
by   Edoardo Manino, et al.
0

Neural networks are a powerful class of non-linear functions. However, their black-box nature makes it difficult to explain their behaviour and certify their safety. Abstraction techniques address this challenge by transforming the neural network into a simpler, over-approximated function. Unfortunately, existing abstraction techniques are slack, which limits their applicability to small local regions of the input domain. In this paper, we propose Global Interval Neural Network Abstractions with Center-Exact Reconstruction (GINNACER). Our novel abstraction technique produces sound over-approximation bounds over the whole input domain while guaranteeing exact reconstructions for any given local input. Our experiments show that GINNACER is several orders of magnitude tighter than state-of-the-art global abstraction techniques, while being competitive with local ones.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
07/28/2023

On the Trade-off Between Efficiency and Precision of Neural Abstraction

Neural abstractions have been recently introduced as formal approximatio...
research
05/17/2021

DISCO Verification: Division of Input Space into COnvex polytopes for neural network verification

The impressive results of modern neural networks partly come from their ...
research
01/27/2023

Neural Abstractions

We present a novel method for the safety verification of nonlinear dynam...
research
04/11/2022

PolyARBerNN: A Neural Network Guided Solver and Optimizer for Bounded Polynomial Inequalities

Constraints solvers play a significant role in the analysis, synthesis, ...
research
06/03/2021

Towards a Mathematical Theory of Abstraction

While the utility of well-chosen abstractions for understanding and pred...
research
04/09/2018

Deep Painterly Harmonization

Copying an element from a photo and pasting it into a painting is a chal...
research
02/14/2012

A temporally abstracted Viterbi algorithm

Hierarchical problem abstraction, when applicable, may offer exponential...

Please sign up or login with your details

Forgot password? Click here to reset