CEREBRA-AI project bid

 

 

 

CEREBRA-AI: A unified framework for continual, explainable and embodied General-Purpose Artificial Intelligence

Horizon Europe Call Details: Topic: HORIZON-CL4-2025-04-DIGITAL-EMERGING-07: Enhanced Learning Strategies for General Purpose AI: Advancing GenAI4EU (RIA) (AI/Data/Robotics Partnership)

Several members of the Network have joined the CEREBRA-AI project consortia (project bid submitted in September 2025).

The CEREBRA-AI project proposes a unified framework for developing Continual, Explainable, and Embodied General-Purpose AI (GPAI) aligned with European values of safety, transparency, and human-centric innovation. CEREBRA-AI responds to key shortcomings of current large AI models—limited adaptability, weak reasoning, and a lack of self-awareness—by integrating neuro-symbolic reasoning, continual learning, and embodied interaction. Its overarching goal is to advance AI that can learn continually, reason logically, and explain its decisions, supporting reliable deployment across sensitive sectors such as healthcare.

Within this framework, the personalised healthcare use case is one of five real-world validation domains. It will focus on developing a clinical decision-support prototype capable of adapting to evolving medical data and reasoning over patient-specific knowledge. By embedding a hybrid neuro-symbolic core, the system will combine data-driven pattern recognition with explicit clinical knowledge graphs, enabling explainable and causally grounded diagnostic insights. This approach directly addresses one of the most pressing challenges in medical AI: ensuring that automated systems can justify their recommendations transparently and remain robust under changing clinical conditions.

CEREBRA-AI’s healthcare demonstrator will test the project’s key innovations — continual learning for updating medical knowledge, causal and counterfactual reasoning for explainability, and embodied or relational learning for contextual understanding of health data. The demonstrator will provide a pathway to trustworthy, adaptive clinical AI, capable of supporting clinicians rather than replacing them. Validation activities will include collaboration with healthcare partners to evaluate decision quality, reliability, and compliance with the EU AI Act.

Overall, CEREBRA-AI positions healthcare as a critical proving ground for next-generation, trustworthy GPAI. By integrating continual adaptation, neuro-symbolic reasoning, and explainability by design, the project aims to set new standards for safe, transparent, and human-aligned AI in medicine, demonstrating how European research can lead globally in ethical and high-impact AI innovation.