Data and AI Technologies

The Data and AI Technologies Task Force uses the Strategic Research and Innovation Agenda (SRIA) and Strategic Research Agenda (SRA) as references to fuel discussions with regards to the European data and AI ecosystem.

Despite the growing importance of data and its wide availability as a key technical driver for Artificial Intelligence and data-driven innovation, significant challenges at both a technical and legal/procedural level still remain. These are listed below:

• The lack of trusted technical systems that enable security and privacy of the data;
• The lack of interoperability of the data and formats;
• The lack of or limited interpretability of AI models (explainable AI);
• Bias in the data and others;
• Guidance and recommendations on Best Practices to use data and AI technologies in industry

The goals of this Task Force are to provide information on such challenges and potential solutions while giving recommendations for future actions in the field of data and AI technologies. Final aim is the development and deployment of data intensive and intelligent services, driven by public-private collaborations, enabling their industry adoption.

The task force will empower industry to leverage AI and data technologies effectively by providing expert guidance and recommendations to innovate responsibly and sustainably, as well as advocating for trusted and secure data practices.

The BDVA Task Force Data and AI Technologies is currently focused on: 

– Usage and adoption of AI and data tools and technologies by industry, assessing existing AI and data solutions in and for industry and understanding the level of adoption and deployment of those tools and technologies

– Key technical topics in Data and AI technologies, such as:   

• New Big Data Architectures – New paradigms in big data architectures and tools, to cope with extreme data requirements and to facilitate the connection to, among others, data spaces and LLMs.
• Distributed AI – Focus on decentralization, real-time decision-making, and collaboration, covering aspects as edge AI, TinyML, federated learning or swarming learning 
• AI Agents – As autonomous system that can independently perceive its environment, make decisions and execute actions to achieve specific goals 
• Frugal AI – Including the design, development and deployment of AI systems that utilize minimal resources to efficiently achieve desired outcomes through more efficient and less resource demanding techniques 
• Human centric AI – Keeping the human in the loop, both in the training and the application phases of the models, enabling a loop where humans provide feedback and suggestions to algorithms 
• Hybrid / neuro-symbolic AI – Combines symbolic reasoning (rules, logic, structured knowledge) with deep learning (data-driven pattern recognition) 
• Synthetic data generation – Generate synthetically high-quality (accurate, relevant & fit-for-purpose) data, either by simulation or AI driven generation (GenAI) 
• AI-ready data / data centric AI – Techniques to prepare data specially for their use by AI applications