[PlanetKR] [CfP] Workshop on Logical Methods for Neural Network Analysis (LogicNN) @ FLoC 2026

[Nicolas Troquard]

See details at
https://perso.ens-lyon.fr/francois.schwarzentruber/research/events/logicnn2026/


# Important dates

- Paper submission: April 27, 2026
- Author notification: June 2, 2026
- Camera-ready submission: June 23, 2026
- Date of workshop: July 25, 2026 (full-day)

# Description

Neural networks power critical applications, from autonomous vehicles to
medical diagnostics, but their opacity pose significant safety and trust
challenges. Formal logical frameworks offer a rigorous path to understand,
validate, and trust these models. This workshop aims to bridge the gap
between logic and neural network analysis.

We invite contributions that demonstrate how logical frameworks can address
core challenges in neural‑network analysis. Topics of interest include the
expressivity of logical formalisms for neural networks, the formal
verification of safety‑critical properties, the development of
interpretation mechanisms grounded in logic, and the assessment of the
computational complexity of model behavior.

- Expressivity and translation: Can logical formalisms accurately represent
neural network behavior? We seek submissions exploring translation of
neural architectures (CNNs, GNNs, RNNs, transformers) into logical
specifications or vice versa.
- Safety verification: Can logic provide formal guarantees of neuron
networks safety-critical properties?
- Logical interpretation: Can logic help building human-understandable
explanations grounded in logic?
- Decidability and computational complexity: Can logic help answer
foundational questions about the tractability of neural network analysis?

# Submissions

Submissions can be in three categories:

- Regular paper: at least 10 pages + bibliography
- Short paper: 5-9 pages + bibliography
- Abstract of already published paper: 1-3 pages + bibliography

The proceedings shall be submitted to CEUR-WS.org for online publication.
Authors of regular and short papers may opt out of having their work
included in the proceedings. Abstracts of already published papers will not
be included in the proceedings.


See details at
https://perso.ens-lyon.fr/francois.schwarzentruber/research/events/logicnn2026/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://kr.org/pipermail/planetkr/attachments/20260208/00990b3c/attachment.htm>