Pedro Porto Buarque de Gusmão is a Lecturer in Computer Science at the University of Surrey, affiliated with the Computer Science Research Centre within the School of Computer Science and Electronic Engineering. His research focuses on Computer Vision, Navigation under adverse conditions, Sensor Fusion, and Distributed Machine Learning. He leads projects in federated learning, autonomous systems, and robotic navigation, contributing to frameworks like Flower, a widely used federated learning library. His work emphasizes advancing federated learning techniques for privacy-preserving distributed systems, with notable achievements including winning the First Prize in the NIST Privacy-Enhancing Technologies Prize Challenge (2023). His publications span conferences like ICCV, USENIX Security, and IEEE Transactions, addressing topics such as federated learning optimization, 3D lane detection, and thermal-inertial SLAM. Key contributions include developing methods like L-DAWA for federated visual learning and FedVal for federated model validation. His research also integrates multimodal sensor data for autonomous navigation systems, leveraging mmWave radar and thermal-inertial fusion. He actively contributes to open-source projects and collaborates with industry on privacy-aware machine learning solutions.
