NeuroSymbolic AI Unpacked: What it is and how HumAIne puts it to work

HumAIne is an EU-funded project developing a novel Operating System for Human-AI collaboration (HumAIne OS), designed to support advanced decision-making in dynamic, unstructured environments across sectors such as healthcare, manufacturing and smart cities. By combining Active Learning, NeuroSymbolic AI, Swarm Learning and eXplainable AI, the HumAIne OS aims to deliver human-centric AI systems that outperform both standalone AI and humans working alone.

In this webinar, we focus on NeuroSymbolic AI (NSAI) and its applications within Smart Healthcare with a focus on the Diabetes pilot. NSAI combines the pattern-recognition power of neural networks with structured domain knowledge in the form of logical rules and expert constraints. This hybrid approach addresses key challenges in healthcare AI: explainability, robustness, trust and data scarcity.

The session will:

• Introduce NeuroSymbolic AI and its core concepts.
• Present the HumAIne NSAI architecture and training approach.
• Show how NSAI is applied to diabetes prediction using the Healthentia dataset.
• Demonstrate NSAI-driven explainability and recommendation generation.

Attendees will see how medical expertise, behavioral guidelines and patient data can be encoded as symbolic rules that shape model training, resulting in more transparent predictions and clinically meaningful advice—while ensuring that final decisions remain with healthcare professionals.

Who should attend?

• Healthcare professionals – Interested in AI-supported monitoring, prediction, and recommendations.
• Digital health providers – Exploring how NSAI integrates into platforms like Healthentia.- Data scientists & ML engineers – Working with rules, explainability, and hybrid neuro-symbolic models.
• Researchers & students – Studying NSAI, XAI, or human-in-the-loop AI methods.
• Decision-makers & policy stakeholders – Evaluating trustworthy, explainable AI for healthcare contexts.
• And anyone applying AI in real-world domains – e.g., manufacturing, finance, logistics, safety-critical systems, where combining data with expert rules can improve reliability and transparency.