US Patent Application 18016468. WATERMARK PROTECTION OF ARTIFICIAL INTELLIGENCE MODEL simplified abstract
Contents
WATERMARK PROTECTION OF ARTIFICIAL INTELLIGENCE MODEL
Organization Name
Telefonaktiebolaget LM Ericsson (publ)
Inventor(s)
Akshara Kannan of CHENNAI (IN)
WATERMARK PROTECTION OF ARTIFICIAL INTELLIGENCE MODEL - A simplified explanation of the abstract
- This abstract for appeared for US patent application number 18016468 Titled 'WATERMARK PROTECTION OF ARTIFICIAL INTELLIGENCE MODEL'
Simplified Explanation
This abstract describes a computer-based method for protecting an artificial intelligence (AI) model from tampering. The method involves determining when the AI model has reached a stable state, known as convergence. Once convergence is detected, a set of baseline parameters for the AI model is identified. These baseline parameters serve as a reference point for the model's behavior. To protect the AI model, a watermark is generated by applying certain transformations to each baseline parameter. This watermark contains a value that is external to the AI model itself. The purpose of this method is to safeguard the AI model from unauthorized modifications or tampering.
Original Abstract Submitted
A computer-implemented model for protecting an artificial intelligence (AI) model from tampering is provided. The method includes determining a convergence of the AI model. The method further includes, responsive to the determining, identifying a set of baseline parameters of the converged AI model. The method further includes generating a first watermark for the converged AI model based on applying one or more transformations to each baseline parameter from the set of baseline parameters, wherein the first watermark comprises a value external to the converged AI model.
