This Python software has been developed as part of the research presented in the scientific paper titled “Dual-Locking Learned AI Models: A PIN-Based Sparse QIM Watermarking and Adaptive Index Permutation Approach.” The operation of the software functions requires downloading AI models that, due to copyright protection, are not included in this repository.

If you find this software useful, please cite the original scientific paper. The citation format is provided below:

Iva Vasic, Jesús Muñoz-Cádiz, and Bata Vasic, “Dual-Locking Learned AI Models: A PIN-Based Sparse QIM Watermarking and Adaptive Index Permutation Approach,” IEEE Transactions on Artificial Intelligence, Vol. XX, No. X, November 2025. DOI: 10.1109/TAI.2025.3636862

This document currently serves as a brief and minimal tutorial whose purpose is to provide an open link to the Python software. It will be expanded and explained in detail soon after the paper is published.

For easier navigation through the files, please note that all files whose names begin with the letter "A" are main scripts that contain function calls to the remaining modules:

• NN for neural network models,

• CNN for convolutional network models, and

• TRF for locking and unlocking transformer-based large language models.

Thank you in advance for your interest in our software, for considering citation, and above all for your willingness to explore this fascinating research domain. We hope you will contribute new ideas and improved results that advance this algorithm even further.