DEEPSPLIT: An Efficient Splitting Method for Neural Network Verification via Indirect Effect Analysis

Henriksen, Patrick; Lomuscio, Alessio

doi:10.24963/ijcai.2021/351

DEEPSPLIT: An Efficient Splitting Method for Neural Network Verification via Indirect Effect Analysis

Patrick Henriksen, Alessio Lomuscio

Download BibTeX Cite

Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence. Main Track. Pages 2549-2555.

https://doi.org/10.24963/ijcai.2021/351

Keywords

Machine Learning: Deep Learning
Multidisciplinary Topics and Applications: Validation and Verification

Abstract

We propose a novel, complete algorithm for the verification and analysis of feed-forward, ReLU-based neural networks. The algorithm, based on symbolic interval propagation, introduces a new method for determining split-nodes which evaluates the indirect effect that splitting has on the relaxations of successor nodes. We combine this with a new efficient linear-programming encoding of the splitting constraints to further improve the algorithm’s performance. The resulting implementation, DeepSplit, achieved speedups of 1–2 orders of magnitude and 21-34% fewer timeouts when compared to the current SoA toolkits.

Patrick Henriksen, and Alessio Lomuscio. 2021. DEEPSPLIT: An Efficient Splitting Method for Neural Network Verification via Indirect Effect Analysis. In Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence, IJCAI-21. IJCAI, 2549–2555. https://doi.org/10.24963/ijcai.2021/351

@inproceedings{ijcai2021p351,
  title     = {DEEPSPLIT: An Efficient Splitting Method for Neural Network Verification via Indirect Effect Analysis},
  author    = {Henriksen, Patrick and Lomuscio, Alessio},
  booktitle = {Proceedings of the Thirtieth International Joint Conference on
               Artificial Intelligence, {IJCAI-21}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},
  editor    = {Prof. Zhi-Hua Zhou},
  pages     = {2549--2555},
  year      = {2021},
  month     = {8},
  note      = {Main Track},
  doi       = {10.24963/ijcai.2021/351},
  url       = {https://doi.org/10.24963/ijcai.2021/351},
}

TY  - CPAPER
TI  - DEEPSPLIT: An Efficient Splitting Method for Neural Network Verification via Indirect Effect Analysis
AU  - Henriksen, Patrick
AU  - Lomuscio, Alessio
T2  - Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence, IJCAI-21
PB  - International Joint Conferences on Artificial Intelligence Organization
PY  - 2021
SP  - 2549
EP  - 2555
DO  - 10.24963/ijcai.2021/351
UR  - https://doi.org/10.24963/ijcai.2021/351
ER  -