Kirill Yankov
Papers
1
Total Citations
2
H-Index
1
About
Kirill Yankov is a researcher advancing the frontier of robot skill generalization through the integration of reinforcement learning and probabilistic modeling. His most-cited work, "Robot Skill Generalization via Keypoint Integrated Soft Actor-Critic Gaussian Mixture Models" (2024), introduces a novel framework that combines keypoint-based representations with the Soft Actor-Critic algorithm and Gaussian Mixture Models. This approach enables robots to learn and adapt manipulation skills across varying task configurations, addressing a core challenge in robotics: transferring learned behaviors to new, unseen scenarios without extensive retraining. By leveraging keypoints to capture task-relevant spatial features, Yankov’s method enhances both sample efficiency and generalization capability, offering a practical pathway toward more versatile autonomous systems. Though his publication record is early-stage, this work has already garnered attention for its elegant synthesis of model-based and model-free techniques. Yankov’s research sits at the intersection of robot learning, computer vision, and control, with potential applications in industrial automation and assistive robotics. His contributions reflect a growing trend toward data-efficient, generalizable skill acquisition—a critical step for deploying robots in unstructured, real-world environments.
Research Focus
Key Achievements
Top Papers
- 1