Workshops
Agnieszka Wykowska
Principal Investigator, Head of Unit “Social Cognition in Human-Robot Interaction" Italian Institute of Technology.From basic research in human-robot interaction to social robots for child healthcare
Abstract: One of the key foci of social robotics is application to child healthcare. Social robots have been developed and tested not only for general educational applications but also for more focused areas, such as robot-assisted training protocols for autism. It is indeed autism where social robots bring most promise of a large positive impact. This is presumably due to the enthusiasm and particular interest that children diagnosed with autism have for technology in general, and robots specifically. In the first part of the tutorial, I will demonstrate how one can develop efficient robot-assisted training protocols for children diagnosed with autism, based on knowledge and results from fundamental research in human-robot interaction. I will provide examples from our lab where we have implemented a program of robot-assisted training of socio-cognitive skills for children diagnosed with autism. Our results show a substantial improvement in the key cognitive mechanisms that are at the core of social cognition and fluent social interaction. The second part of the tutorial will be more interactive, where participants will engage in group work and will try to develop a new robot-assisted training protocol for children with disabilities, based on some methods used in basic research that I will provide.
Christian Barceló
Software Developer EkumenSimulation-as-Code: A Guide to Synthetic Data Generation with Isaac Sim
Abstract: High-quality, labeled data is the fuel for modern AI, yet its acquisition remains a critical bottleneck. This tutorial dives into a scalable solution using NVIDIA Isaac Sim, treating the simulator not as an interactive tool, but as a powerful, Python-native rendering and physics engine. We adopt a "simulation-as-code" methodology, building a data generation pipeline entirely through scripts. Attendees will use a provided code repository and environment to write Python scripts that define every aspect of a virtual scene—from loading assets to positioning cameras. We will then integrate the high-level Replicator API to orchestrate sophisticated Domain Randomization, procedurally varying textures, poses, and lighting. The organizers will run these scripts live, demonstrating how to extract perfect ground-truth data and execute the entire pipeline in headless mode. Participants will leave with a powerful, working script and a production-oriented mindset for generating datasets at scale, all without needing specialized local hardware. To participate, attendees are required to bring their own personal computer.
Giorgio Metta
Istituto Italiano di Tecnologia Scientifica DirectorTO BE DETERMINED
Abstract:
Gustavo J. G. Lahr & Team
Assistant professor, Hospital Israelita Albert EinsteinInstituto Israelita de Ensino e Pesquisa.
Robotic Fault-Detection Strategies for Complex Industrial Contexts
Abstract: Robotic platforms are playing an increasingly vital role in the inspection and maintenance of complex industrial assets. Equipped with mobile bases and a diverse suite of sensors, including vision, thermography, ultrasound, and vibration monitors, these systems can autonomously navigate large facilities, gather high-resolution data, and identify early-stage faults long before they escalate into costly failures. In several industries, robotic solutions have begun to demonstrate their value; however, different sectors face challenges in their deployment. Harsh environmental conditions, heterogeneous legacy equipment, stringent safety regulations, and the need for real-time decision support all hinder the seamless integration of autonomous inspection systems. Moreover, advanced data-fusion algorithms and AI-driven anomaly-detection methods, while promising in the lab, often struggle to scale reliably in the field without close industry feedback. This full-day workshop will convene experts from industry and academia to survey current implementations, dissect technical and organizational bottlenecks, and chart a collaborative roadmap for robot-centered failure detection. Through a combination of keynote talks, interactive presentations, panel discussions, and poster sessions, participants will leave with actionable insights into deploying mobile inspection robots, integrating advanced sensing modalities, and leveraging AI for predictive maintenance—laying the groundwork for new partnerships throughout Latin America and beyond.