Samuel Wilson

Dr Samuel Wilson is a researcher at the QUT Centre for Robotics (QCR), holding a PhD in Computer Science (2024). His work focuses on uncertainty estimation in deep-learning-based visual perception for robotics, leveraging platform-specific characteristics like egocentric flow and embodiment. Supervised by Dr Feras Dayoub and Assoc Prof Niko Sünderhauf, he transitioned to PhD research after observing fragility in vision models during his Honours project (2019).

Education:

• Bachelor’s in Computer Science, First Class Honours (2019)

Research Interests: Enhancing robotic perception reliability through uncertainty quantification, with emphasis on real-world robustness and integration of robotics-specific sensor data. His methods emphasize practical deployment challenges in autonomous systems.

Publications Trends: Recent work concentrates on out-of-distribution detection and feature-based approaches to improve model reliability in unstructured environments, reflecting a shift toward safety-critical robotics applications.

Labs/Teams: Active contributor to QCR’s research programs in autonomous systems and computer vision.