Speaker: Zhaoxing Deng
Title: Automated Skill Evaluation of Bronchoscopy Operators Using Geometric Feature Analysis
Abstract: Bronchoscopy is a critical procedure for diagnosing and treating lung conditions, but current training assessments are subjective and time-consuming. More objective skill assessment methods are needed to ensure clinician competence. We propose an automatic skill assessment method to classify bronchoscope operators as experts or novices. Pose data from an electromagnetic tracker was collected during bronchoscopy on a lung model. We defined geometric features that capture key aspects of operator performance and used them to train a classifier distinguishing experts from novices. Our results show that expert operators adhere more closely to the geometric features, with smaller deviations and greater consistency, while novice operators exhibit higher variance.
Speaker: Xiaofeng Mao
Title: Learning Long-Horizon Robot Manipulation Skills via Privileged Action
Abstract: Long-horizon contact-rich tasks are challenging to learn with reinforcement learning, due to ineffective exploration of high-dimensional state spaces with sparse rewards. The learning process often gets stuck in local optimum and demands task-specific reward fine-tuning for complex scenarios. In this work, we propose a structured framework that leverages privileged actions with curriculum learning, enabling the policy to efficiently acquire long-horizon skills without relying on extensive reward engineering or reference trajectories. Specifically, we use privileged actions in simulation with a general training procedure that would be infeasible to implement in real-world scenarios. These privileges include relaxed constraints and virtual forces that enhance interaction and exploration with objects. Our results successfully achieve complex multi-stage long-horizon tasks that naturally combine non-prehensile manipulation with grasping to lift objects from non-graspable poses. We demonstrate generality by maintaining a parsimonious reward structure and showing convergence to diverse and robust behaviors across various environments. Our approach outperforms state-of-the-art methods in these tasks, converging to solutions where others fail.