HORIZON-CL4-2025-04-DIGITAL-EMERGING-07: Enhanced Learning Strategies for General Purpose AI: Advancing GenAI4EU

We offer our expertise in AI to endow GPAI models with abilities of "long-term planning, and seamless adaptation to dynamic and non-stationary environments", and to seek "the development of groundbreaking GPAI models that combine self-supervised learning with complementary learning strategies [such as] hybrid learning". This could be done, for instance, by integrating goal reasoning (the process by which intelligent agents continually reason about the goals they are pursuing, which may lead to goal change and adaptation to complex dynamic environments), as we are currently doing in a project funded by ONRG, or by integrating probabilistic program synthesis of perturbations to domain transition models, as we did in a project funded by DARPA ( https://authors.elsevier.com/c/1l2~M-c5jN1g ).