Ruan, Wenjie, Huang, Xiaowei 
ORCID: 0000-0001-6267-0366 and Kwiatkowska, Marta
(2018)
Reachability Analysis of Deep Neural Networks with Provable Guarantees.
In: 27th International Joint Conference on Artificial Intelligence, 2018-7-13 - 2018-7-19, Stockholm, Sweden.
![[img]](https://livrepository.liverpool.ac.uk/style/images/fileicons/text.png) |
Text
1805.02242.pdf - Published version Download (1MB) |
Abstract
Verifying correctness of deep neural networks (DNNs) is challenging. We study a generic reachability problem for feed-forward DNNs which, for a given set of inputs to the network and a Lipschitz-continuous function over its outputs, computes the lower and upper bound on the function values. Because the network and the function are Lipschitz continuous, all values in the interval between the lower and upper bound are reachable. We show how to obtain the safety verification problem, the output range analysis problem and a robustness measure by instantiating the reachability problem. We present a novel algorithm based on adaptive nested optimisation to solve the reachability problem. The technique has been implemented and evaluated on a range of DNNs, demonstrating its efficiency, scalability and ability to handle a broader class of networks than state-of-the-art verification approaches.
| Item Type: | Conference or Workshop Item (Unspecified) |
|---|---|
| Additional Information: | This is the long version of the conference paper accepted in IJCAI-2018. Github: https://github.com/TrustAI/DeepGO |
| Uncontrolled Keywords: | cs.LG, cs.LG, cs.CV, stat.ML |
| Depositing User: | Symplectic Admin |
| Date Deposited: | 09 May 2018 09:13 |
| Last Modified: | 19 Jan 2023 06:34 |
| Related URLs: | |
| URI: | https://livrepository.liverpool.ac.uk/id/eprint/3021099 |
Share
Share
