The 6th International Verification of Neural Networks Competition (VNN-COMP 2025): Summary and Results
Konstantin Kaulen, Tobias Ladner, Stanley Bak, Christopher Brix, Hai Duong, Thomas Flinkow, Taylor T. Johnson, Lukas Koller, Edoardo Manino, ThanhVu H Nguyen, Haoze Wu · 2025
Read The 6th International Verification of Neural Networks Competition (VNN-COMP 2025): Summary and Results by Konstantin Kaulen, Tobias Ladner, Stanley Bak, Christopher Brix, Hai Duong, Thomas Flinkow, Taylor T. Johnson, Lukas Koller, Edoardo Manino, ThanhVu H Nguyen, Haoze Wu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2512-19007.txt.
Introduction
Konstantin Kaulen and co-authors sum up VNN-COMP 2025, the sixth yearly contest for tools that verify neural networks. Verification proves a network meets a stated property, rather than just testing sample inputs. The contest ran inside a symposium on AI verification held with a major computer-aided verification conference.
Eight teams entered, and their tools ran on shared benchmark sets with standard file formats and equal-cost cloud hardware. Settings were locked before the final test sets were revealed. The report covers the rules, sixteen regular and nine extended benchmarks, the results, and lessons for making future comparisons fairer and more useful.
Catalog record
| CREATOR | Konstantin Kaulen, Tobias Ladner, Stanley Bak, Christopher Brix, Hai Duong, Thomas Flinkow, Taylor T. Johnson, Lukas Koller, Edoardo Manino, ThanhVu H Nguyen, Haoze Wu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 443 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2512-19007