Generating Heuristics Using Weakly-Supervised Machine Learning and Digital Twin Simulations

The EPO published patent application EP2024201122A1 for a method of generating heuristics using weakly-supervised machine learning combined with digital twin simulations. The application is classified under G06N 5/01 (knowledge engineering), G06N 5/022 (knowledge models), and G06N 20/00 (machine learning). The invention relates to AI systems that use digital twin technology alongside weakly-supervised learning to develop heuristic-based solutions.\n\nTechnology companies and research institutions developing similar AI methodologies should review this publication to assess potential patent landscape implications. Companies actively filing patents in machine learning, digital twin, or hybrid AI simulation technologies may need to conduct freedom-to-operate analyses. The patent designation covers all EPO member states, making it relevant for entities seeking European patent protection in these technical domains.