Alessio Xompero

Alessio Xompero is a Research Fellow at Queen Mary University of London's School of Electronic Engineering and Computer Science. His work focuses on robotics, computer vision, and human-robot interaction, particularly in object affordance understanding, multimodal perception, and safe human-robot collaboration. He develops algorithms for robotic manipulation, perception under occlusion, and adversarial robustness in AI systems. Xompero contributes to benchmark datasets like CORSMAL and has published extensively on topics including container property prediction, audio-visual classification, and view-overlap recognition across cameras.

His research integrates machine learning with robotics applications, addressing challenges such as real-time grasping, privacy-aware AI models, and sensor fusion. Notable projects include the mixed-reality dataset for 6D pose estimation and work on adversarial attacks/defenses in visual perception systems. Xompero's contributions advance both theoretical understanding and practical implementations of autonomous systems.

Key themes in his recent work include improving robotic systems' adaptability to unstructured environments, enhancing safety in human-robot interactions, and developing interpretable models for privacy preservation. He actively participates in robotics competitions and benchmark evaluations to validate algorithmic performance under real-world conditions.