Call for Papers

We invite participants from any discipline related to formal verification of machine learning, including but not limited to machine learning, computer-aided verification, programming languages, robotics and control, computer security, and optimization. We invite submissions on any aspect of the formal verification of machine learning, which includes but not limited to:

  • New formal verification methods for different types of machine learning models, including neural networks (NNs), tree-based models, Gaussian process, nonparametric models, and many others;

  • Algorithms for verifying novel or unconventional formal specifications, especially those that arose in new applications of verified machine learning;

  • Scaling up formal verification algorithms for machine learning;

  • Novel model training procedures that produce high assurance models friendly to formal verification;

  • New applications where formal verification is applied, such as applications in robotics, control, aeronautics, computer security and computer systems;

  • Study on the trade-offs between verifiable guarantees, model performance, and model applicability;

  • Design of representative and generic benchmarks and datasets for machine learning verification;

  • Development of user-friendly verification toolkits for machine learning practitioners.

Reviewing will be performed in double-blind, with criteria include:

  • Quality of the methodology and experiments

  • Novelty

  • Relevance

  • Societal impacts

Author Instructions

Submission Deadline: June 3, 2022, Any time on Earth (AoE)

Submission Link: Papers should be submitted to CMT: https://cmt3.research.microsoft.com/WFVML2022

Submission Format: Submitted papers are recommended to have at most 6 pages with unlimited bibliography and appendix, using the WFVML 2022 LaTeX style files. The page limit is not hard, but keeping your paper reasonably short and concise can greatly improve readability. Submissions must be anonymous following ICML requirements and ethics standards. Accepted papers will be hosted on this workshop website but are considered non-archival and can be submitted to other workshops, conferences or journals if their submission policy allows.

Dual submission policy: We allow dual submission with other workshops or conferences. You can also submit papers that have large overlaps with your recently published papers.

Author Notification: Sent

Camera Ready and Video Submission: June 27, 2022 AoE

Best Paper Award🏆: To encourage high-quality submissions, we plan to offer a best paper award (details TBA).

Contact: For technical problems regarding submission, please contact Huan Zhang (huan@huan-zhang.com) and Kaidi Xu (kx46@drexel.edu)