We are pleased to share that Aggelos Psiris, PhD Candidate and Research Associate at Harokopio University of Athens, represented the TORNADO project at the Multimodal Anomaly Detection Workshop 2026 (MMAD 2026), hosted at Aalborg University | February 2, 2026 · Aalborg, Denmark.
About MMAD 2026
The Multimodal Anomaly Detection Workshop 2026 (MMAD 2026) is an international scientific event dedicated to the latest advances in anomaly detection across multiple data modalities, including vision, audio, and sensor data. The workshop brings together researchers and practitioners to explore cutting-edge methods for detecting rare or unexpected patterns in complex, real-world environments, with strong relevance for industrial applications, AI systems, and smart manufacturing.
TORNADO’s Contribution
During the workshop, Aggelos Psiris presented ongoing research carried out in collaboration with Georgios Th. Papadopoulos, focusing on industrial few-shot defect detection.
The work introduces an innovative approach to learning normality from a very limited number of defect-free images, using masked graph reconstruction techniques. This method enables AI systems to identify anomalies even when only scarce training data is available—an important challenge in real industrial settings where defective samples are rare or costly to obtain.
Impact for TORNADO
This research represents a key scientific outcome of the TORNADO EU project, contributing to the development of robust, data-efficient anomaly detection solutions for next-generation industrial environments.
By advancing few-shot learning techniques, TORNADO supports more reliable quality control processes, reduced waste, and increased efficiency—paving the way for smarter and more sustainable manufacturing systems.
