Generative & Agentic AI

Date: 18 June 2026

Room: CB30B

Chairs: Giovanni Bonetta (FBK), Bernardo Magnini (FBK), Marco Polignano (UniBA), Giovanni Semeraro (UniBa)

Description

Generative Artificial Intelligence (GenAI) represents an advanced category of artificial intelligence, characterized by the ability of its models to generate original content, including text, images, videos, audio tracks, and even software code. This innovative AI paradigm uses deep learning algorithms to simulate the human creative process, finding applications in a multitude of fields, from literary production to biological research, from manufacturing to the financial sector. Recent developments indicate that GenAI is exerting a transformative influence on the economy and the business landscape at an extraordinary pace: projections from the McKinsey Global Institute anticipate an increase in the global economic impact of artificial intelligence ranging from 15% to 40% over the coming decades. GenAI-based tools, such as ChatGPT, DALL-E, DeepSeek, and Claude, are reshaping the operational methods of numerous professions. In parallel, recent research has increasingly explored the integration of generative models within agentic systems, where AI models are capable not only of generating content, but also of reasoning, planning, and interacting with external tools and environments to accomplish complex tasks. These Agentic AI systems extend the capabilities of generative models towards more autonomous and interactive settings, opening new challenges in terms of reliability, control, and evaluation. While generative models already raise important ethical and societal concerns, the emergence of agentic systems further amplifies these challenges, making it crucial to address issues such as bias, reliability, and their impact in real-world decision-making contexts. The workshop’s objective is to bridge the gap between academia and industry, fostering dialogue and reflection on both the technological aspects and the societal impacts that Generative and Agentic AI are bringing about.

Detailed Programme

Times show each contribution's presentation slot. An asterisk marks the presenting author when provided in the programme data.

TimeIDContribution
09:00-09:10#43Preserving Memory, Expanding Creativity: Human-Centered AI Trajectory in Engineering and Music Research at the CSC of Padua UniversityCanazza, Sergio; Fiordelmondo, Alessandro; Giuriati, Sara; Paulon, Cristina; Pretto, Niccolò; Rodà, Antonio; Russo, Alessandro; Spanio, Matteo; Zuccante, Anna
09:10-09:20#107Can LLMs understand LilyPond? A benchmark for symbolic music generation and understandingSpanio, Matteo; Torabi, Mohammad; Poltronieri, Andrea; Rodà, Antonio
09:20-09:30#91Human-in-the-Loop Multimodal LLMs for UX Evaluation of Educational Recommender SystemsStefano, Valtolina; Matamoros, Ricardo Anibal; Epifania, Francesco
09:30-09:40#96When to Intervene? An Agentic Framework for Uncertainty-Triggered Human-in-the-LoopGodinho, Alexandre
09:40-09:50#119From Generative to Agentic AI: Developments, Architectures and Applications Across Domains.Amato, Flora; Benfenati, Domenico; Capuozzo, Salvatore; Cirillo, Egidia; De Filippis, Giovanni Maria; Fonisto, Mattia; Foscardi, Gianfranco; Marassi, Lidia; Marrone, Stafano; Moccardi, Alberto; Moscato, Vincenzo; Patwardhan, Narendraprakash; Rinaldi, Antonio Maria; Russo, Cristiano; Sansone, Carlo; Tommasino, Cristian
09:50-10:00#14Generative and Agentic Artificial Intelligence for Extended Reality Scenario Authoring in Industrial Robotics TrainingMartis, Andrea Giovanni; Reforgiato Recupero, Diego; Riboni, Daniele; Scarpi, Giuseppe; Marras, Mirko
10:00-10:10#20Integrated Generative Artificial Intelligence Workflows for Multi-varietal Low-Resource Language TechnologiesCarta, Salvatore Mario; Concas, Filippo; Fenu, Gianni; Giuliani, Alessandro; Manca, Marco Manolo; Marras, Mirko; Podda, Alessandro Sebastian
10:10-10:20#41Parameter-Efficient Multi-View Proficiency Estimation: From Discriminative Classification to Generative FeedbackBianchi, Edoardo; Liotta, Antonio
10:20-10:30#18Generative Models for Trustworthy Ergonomic FeedbackDibenedetto, Gaetano; Polignano, Marco; Lops, Pasquale
10:30-11:00Coffee Break
11:00-11:10#112Human-Centered AI in Higher Education: Toward a Shared Framework for AI Literacy and Faculty Development in the Age of Generative AIAdorni, Giovanni
11:10-11:20#68Usage of Visual Language Models in a robotic context for unknown environments descriptionSiragusa, Irene; Contino, Salvatore; Picone, Alessandro; Pirrone, Roberto
11:20-11:30#37Estimating the Scale of Closed-Source Large Language Models from API PricingPremoli, Giorgio; Foà, Elena; Bertolotti, Francesco
11:30-11:40#35The Declining Cost of Intelligence: Temporal and Economic Trends in LLM Performance on GPQAFoà, Elena; Premoli, Giorgio; Bertolotti, Francesco
11:40-11:50#62Who Builds the Experts? Identity Framing and Generator Bias in Multi-Agent LLM SystemsBellan, Patrizio; Rampazzo, Mattia; Dragoni, Mauro
11:50-12:00#61Agent Identity EvaluationBellan, Patrizio; Dragoni, Mauro; Sanna, Leonardo; Maruotti, Piergiorgio; Rampazzo, Mattia; Muñoz, Sergio
12:00-12:10#106Divide, Deliberate, Decide: A Multi-Agent Framework for Fine-Grained Egocentric Action RecognitionSottovia, Alessandro; Torcinovich, Alessandro; Lanz, Oswald
12:10-12:20#82A Modular GraphRAG Framework for Grounded Strategic Reasoning with Knowledge Graphs and LLMsSiciliani, Lucia; Ghizzota, Eleonora; Jordan, Alex; Petruzzelli, Alessandro; Spillo, Giuseppe; Zaza, Gianluca; Basile, Pierpaolo; Sola, Davide; Lops, Pasquale; Borioli, Giovanni Scarso; Semeraro, Giovanni
12:20-12:30#54SAI4EO: From Conversational Remote-Sensing Tool Use to Artifact-Grounded Geospatial IntelligenceTaggio, Nicolò; Zeggada, Abdallah; Monaco, Domenico; Nicastro, Nicola; Oranger, Edoardo; Musacchio, Elio; Semeraro, Giovanni
12:30-12:40#57A Transparent AI Assistant for Course LearningAbatescianni, Lorenzo; Creanza, Maria; Lops, Pasquale; Pinto, Nicolas; Polignano, Marco; Semeraro, Giovanni
12:40-12:50#113Generative AI in Radiology: Synthesis, Enhancement, and Unified ModelingAksu, Fatih; Caragliano, Alice Natalina; Caruso, Camillo Maria; Di Feola, Francesco; Iele, Irene; Mantegna, Massimiliano; Molino, Daniele; Mulero Ayllón, Elena; Pesci, Alessandro; Pompilio, Ludovica; Rofena, Aurora; Romoli, Giulia; Ruffini, Filippo; Salmè, Marco; Sicilia, Rosa; Guarrasi, Valerio; Tortora, Matteo; Soda, Paolo
12:50-13:00#78Actor–Critic LLM Architecture for Validation-Driven Code Remediation: A Cross-Model Benchmark and Production StudyZuccari, Lorenzo; Saqib, Najmus; Altobelli, Aldo; Previtali, Fabio
13:00-13:10#85Chain-to-Cognition: From Blockchain-Certified Events to Agentic Supply Chain Digital TwinsVitale, Ciro; Catalano, Pietro; Moscariello, Salvatore; Perillo, Pasquale; Ciaramella, Angelo; Di Capua, Michele; De Prisco, Aniello

Topics of Interest

The topics of interest for the workshop include, but are not limited to:

  • Generative AI for building symbiotic systems;
  • Generative AI, creativity, and art;
  • Generative AI and discovery and recommendation systems;
  • Development of conversational systems using generative AI;
  • The impact of generative AI on software development;
  • The impact of generative AI on education;
  • Fairness of generative AI models;
  • Generative AI for medicine;
  • Personalization in the retail sector using generative AI;
  • Generative AI for the entertainment industry;
  • Large Language Models (LLMs) and Large Multimodal Models (LMMs);
  • Large Concept Models
  • Human-AI collaboration and human-in-the-loop agentic systems;
  • Agentic AI and autonomous decision-making systems;
  • Evaluation and benchmarking of agentic AI systems;
  • Multi-agent systems;