Finalizing VNN-COMP'25 report

[Andrew Wu]

Dear VNN community,

Thank you for contributing to the VNN-COMP'25 report. We have mostly finalized the report and plan to upload it to ArXiv by the end of this month. If there are changes you would like to make, please do so by the end of this week: https://www.overleaf.com/4731569512prbbckmrymwq#907ec5

We would also like to make you aware that there was a discussion about the nuemrical robustness of the counterexamples found by the verifiers, when we were preparing the report. In a nutshell, there were cases where a particular input constituted a valid counter-example on one hardware but would not lead to a violation on another hardware. For this reason, we decided to compute and report two sets of results, one with zero tolerance for checking the validity of the counter-examples, the other with a small tolerance. In terms of ranking, it only swapped the order of #5 and #6.  Please refer to Section 5 of the report for further information, and the conclusion section for additional discussions. If you have any thoughts or questions about this point, please don't hesitate to reach out to us.

Thank you for your contributions and support.

Best wishes,

The VNN-COMP 2025 Organizers