Adrien Chan-Hon-Tong
Université Paris-Saclay, Office National d'Études et de Recherches Aérospatiales
Papers
2
Total Citations
6
H-Index
2
About
Adrien Chan-Hon-Tong is a researcher focused on bridging the gap between simulation and reality in robotics, with key contributions in deep reinforcement learning and computer vision. His most cited work, "Sim-to-Real Transfer with Incremental Environment Complexity for Reinforcement Learning of Depth-based Robot Navigation" (2020, 4 citations), tackles the critical challenge of transferring learned navigation policies from simulated environments to real-world robots. By introducing incremental environment complexity during training, he demonstrates a practical method to improve model robustness and sample efficiency, addressing a fundamental bottleneck in model-free control theory. In his earlier work, "Pertinence of Video for Single Image Deep Network" (2017, 2 citations), Chan-Hon-Tong challenges conventional wisdom by showing that training single-image deep networks on all video frames—rather than just key frames—can significantly boost performance on medium-sized datasets. This insight has implications for data-efficient learning in computer vision. His research advances the practical deployment of autonomous systems, making him a notable contributor to the fields of sim-to-real transfer and visual learning for robotics.
Research Focus
Key Achievements
Top Papers
- 1
- 2Pertinence of Video for Single Image Deep Network2 citations · 2017