A new framework for experimentation-driven analytics
Extreme data characteristics represent a challenge for advanced data-driven analytics and decision-making in critical domains such as crisis management, predictive maintenance, mobility, public safety and cyber-security. Data-driven insights must be timely, accurate, precise, fit-for-purpose and reliable, considering and learning from user intents and preferences. The EU-funded ExtremeXP project will create a next-generation decision support framework that integrates novel research from big data management, machine learning, visual analytics, explainable ΑΙ, decentralised trust, and knowledge engineering. The framework will aim at optimising the properties of complex analytics processes (e.g. accuracy, time-to-answer, specificity, recall, precision, resource consumption) by associating different user profiles with computation variants, promoting a human-centered, experimentation-based approach to AI and complex analytics. The project will perform five pilot demonstrations.