DEUSTO FOR KNOWLEDGE. D4K

D4K focuses its research on applied computational AI with real-world impact. After more than 20 years of knowledge transfer, D4K is consolidating a next-generation AI for Society that combines scientific excellence, effective transfer and university–industry cooperation, aligned with RIS3 Euskadi and with a leading focus on Smart Industry. At D4K, we are advancing our agenda around four Next-Generation AI axes that respond to the new challenges facing AI itself and to those of Smart Industry, at a high-value intersection of knowledge: - QUANTUM AI (QAI/QML). Quantum-classical hybrid optimisation and learning (QAI, QML, QNN, quantum kernels) with quantum backend and classical fallback; quantum vision for inspection/metrology with distillation to edge/PLC/IIoT models; quantum cybersecurity and PQC integrated into MLOps with end-to-end lineage. - PHYSICAL AI (perception-action). Multimodal models (vision-language-action) with industrial semantics via Knowledge Graphs and GraphRAG; Collaborative Robotics (HRC) with runtime assurance; unstructured bin picking and 6D metrology with sim2real, domain randomisation and control (including fractional approaches). - COLLECTIVE AI (multi-agent/edge-native swarm). Decentralised coordination (CTDE, consensus, auctions) for intralogistics and cells; fleet MLOps; semantics at scale with KG/GraphRAG for mass cloning across stations and plants. - CONSCIOUS AI (governance, explainability, traceability, sustainability). Traceable XAI, data-weight-parameter lineage, and uncertainty metrics (ECE, Drift, OoD); HITL with thresholds/roles and runtime monitors; Green-MLOps with ultra-high-speed telemetry. All of this in pursuit of our continuing mission to serve Society, moving towards greater sustainability, competitiveness and understanding.