Agon Serifi

Papers

Total Citations

H-Index

About

Agon Serifi is a rising researcher at the intersection of computer graphics, robotics, and physics-based character animation. His work focuses on bridging the gap between simulated motion and real-world robotic control, with a particular emphasis on motion generation and robust tracking for physical characters. Serifi’s most influential contribution is the “Robot Motion Diffusion Model” (2024, 13 citations), which introduces a generative framework for producing realistic, physically plausible motion for robotic characters—a key step toward more expressive and adaptable robots. In his follow-up work, “VMP: Versatile Motion Priors” (2024, 10 citations), he tackles the challenge of training a single control policy that can handle diverse and unseen motions, enabling robust deployment on real-world physical robots. This work is notable for its potential to unify motion tracking across varied hardware. Additionally, his “Transformer-Based Neural Augmentation of Robot Simulation Representations” (2023, 3 citations) addresses persistent sim-to-real gaps by using transformer architectures to correct modeling errors like friction. Together, Serifi’s research is laying the groundwork for more fluid, reliable, and generalizable robotic motion, making him a promising voice in the future of embodied AI.