Peer-reviewed research

The collaborative research conducted between Art Recognition and the Department of Cognitive Science & Artificial Intelligence (CSAI) at the University of Tilburg has been published following a rigorous peer-review process. The findings presented in the paper underscore the remarkable potential of Transformers — an advanced neural network architecture originally tailored for extensive language models such as ChatGPT — in the domain of image classification for art authentication. This innovation marks a significant stride in enhancing the reliability of computer-based authentication for artworks, resulting in heightened precision and confidence throughout the authentication process. Another peer-reviewed scientific article, co-authored by Art Recognition and the Department of Mathematical Sciences at the University of Liverpool, has been released. This article demonstrates the capability of our AI system to effectively resist adversarial attacks after incorporating AI-generated images into its training dataset.