Conveners
Artificial Intelligence (AI) - I
- Daniele Bonacorsi (University of Bologna)
Description
During the last decade, Artificial Intelligence (AI) and statistical learning techniques have started to become pervasive in scientific applications, exploring the adoption of novel algorithms, modifying the design principles of application workflows, and impacting the way in which grid and cloud computing services are used by a diverse set of scientific communities. This track aims at discussing problems, solutions and application examples related to this area of research, ranging from R&D activities to production-ready solutions. Topics of interests in this track include: AI-enabled scientific workflows; novel approaches in scientific applications adopting machine learning (ML) and deep learning (DL) techniques; cloud-integrated statistical learning as-a-service solutions; anomaly detection techniques; predictive and prescriptive maintenance; experience with MLOps practices; AI-enabled adaptive simulations; experience on ML/DL models training and inference on different hardware resources for scientific applications.
Commercial software development today happens almost exclusively in agile teams, using either the SAFe or Scrum framework. AI agents can communicate via MCP and take on different roles, including developer, scrum master, product owner and QA roles. The presentation discusses how to set up agile teams of agentic AIs, giving human developers the opportunity to let a group of AIs develop a...
Reliable uncertainty quantification is essential in scientific applications, where predictive results must be supported by a transparent assessment of confidence. Among the many approaches proposed for this purpose, Conformal Prediction (CP) is especially compelling because it offers finite-sample, distribution-free coverage guarantees and can calibrate uncertainty on top of any trained model...
Spatio-temporal data mining is effective for extracting useful information from the occurrence frequencies and patterns of real-world physical phenomena. The author has previously proposed a spatio-temporal and categorical data mining method that not only extracts occurrence frequencies and patterns from spatio-temporal features, but also performs semantic interpretation of relationships...
The shift from conventional computing, networking, and storage to AI-driven scientific discovery (AI4S) calls for a new generation of intelligent infrastructure. In this era of agentic AI, scalable access to models, tools, data, and agents has become critical—serving as essential utilities powering next-generation research. To meet these demands, the HepAI team has developed Qionwu, a core...
