Menu
An integrated platform for Green and Energy-efficient data and model related operations…
RAIDO’s Vision is a comprehensive framework for Trustworthy and Green-AI, offering a comprehensive solution for data and model-related aspects. The platform includes automated data curation methods, data-efficient models, and tools for energy-efficient Green AI. Transparency, explainability, and soundness of optimized AI models and data handling processes are ensured through various XAI methods, decentralized blockchain, feedback-based reinforcement learning, novel KPIs, and visualization techniques. A novel AI orchestrator is introduced to optimize tasks and processes, reducing energy consumption and environmental footprint. The integrated platform will be evaluated through four real-life demonstrators in key application domains.
• O-1: Automatically enhance data quality and perform data augmentation for energy efficient AI
• O-2: Generate large volumes of synthetically generated data with corresponding annotations
• O-3: Optimise learning processes and models without quality degradation
• O-4: Develop AI Orchestrator for creating an optimized dataset & training pipeline tailored to the application in hand
• O-5: Enhance the explainability, fairness, and transparency of the AI models
• O-6: Develop AI framework benchmarking, and progress monitoring and feedback to ensure continuous improvement
• O-7: Optimise and automate the AI E2C pipeline and performance
Power & Energy Grid Management for AI-enabled Optimal Planning
Pilot Site - AYE (Spain) & PPC (Greece)
Technology Providers - AYE, PPC, SID, CERTH, TCD, UBITECH, MINDS
Innovation Providers - SID, AYE, CERTH, UBITECH, AYE, MINDS
The pilot aims to improve energy grid management using AI-enabled optimal planning, addressing challenges in planning due to renewable energy sources. It involves AYE and PPC, along with technology providers like SID, CERTH, TCD, UBITECH, and MINDS. The pilot uses big data from sensors, devices, and appliances to train predictive models and optimize them using the RAIDO platform. The pilot will involve interconnected power plants, power grids, and smart homes, using Digital Twins and simulated data for training and optimization.
Autonomous & AI Powered Monitoring & Operations
Pilot Site - HELD/KRE (UK/Greece, Belgium)
Technology Providers - HELD, KRE, 8BELLS, TCD, UBITECH, QMUL
Innovation Providers - KU, SID, QMUL
This Pilot focuses on precision agriculture using autonomous and AI-powered monitoring and operations in the UK, Greece, and Belgium. It addresses pharmaceutical cannabis cultivation and fungal processes for meat replacement. The pilot uses multispectral photography and AI training models to detect diseases and predict optimal harvesting times. The setup includes IoT devices, RGB-D cameras, and multispectral sensors. The demonstration will use historical data to build digital twins and train AI models.
Digital Health Solutions for Personalised Preventive Pharmacogenetics
Pilot Site - VITO, JESSA (Belgium)
Technology Providers - VITO, TCD, SID, 8BELLS
Innovation Providers - VITO, KU, QMUL, SID
Pilot #3 aims to improve healthcare by implementing an AI system to design personalized preventive pharmacogenetics (PGx) application archetypes. The pilot addresses challenges like diversity in actionable PGx information, socio-economic status, and healthcare professionals' understanding of PGx information. The system uses AI models, bias detection, feedback, and blockchain technology to improve communication, reduce unfairness, and enhance safety and efficacy of drugs. The pilot will involve patients and healthcare professionals in real-life settings.
Industry 5.0 & Bio-based Composites, AI Models for Plant Fibre
Pilot Site - UBFC (France)
Technology Providers - UBITECH, TCD, UBFC, SID, 8BELLS
Innovation Providers - UBFC, MINDS, SID, 8BELLS
Pilot #4 aims to develop Plant Fiber Reinforced Composites (PFRCs) using advanced deep reinforcement learning architectures and human-in-the-loop (HITL) to address high cost and limited properties of biodegradable materials. The pilot will involve data generation/collection, AI V-L models design and training, and optimization and explainable AI (XAI). The goal is to achieve 15x faster mechanical property estimation with over 95% accuracy.
