Papers
13
Total Citations
341
H-Index
8
About
Nicklas Hansen is a rising researcher at the forefront of visual reinforcement learning, robot learning, and generalization in deep RL systems. His work addresses one of the field's most persistent challenges: enabling agents trained in simulated or controlled environments to transfer effectively to novel, real-world settings. His most influential contribution, "Generalization in Reinforcement Learning by Soft Data Augmentation" (2021, 100 citations), introduced a principled approach to data augmentation that balances exploration and optimization stability, becoming a key reference for practitioners in visual RL. Complementing this, his work on self-supervised policy adaptation during deployment (58 citations) demonstrated how agents can continually adjust to environmental shifts without human intervention — a critical capability for real-world deployment. Hansen has also pushed boundaries in robotic manipulation, leveraging transformers to bridge egocentric and third-person views (47 citations), and in quadrupedal locomotion through cross-modal transformer architectures. His more recent explorations into 3D self-supervised representations and visuo-motor world models signal a trajectory toward increasingly capable and generalizable robotic systems. Collectively, Hansen's research has garnered over 330 citations, establishing him as a significant voice in modern embodied AI research.
Research Focus
Key Achievements
Top Papers
- 1Generalization in Reinforcement Learning by Soft Data Augmentation100 citations · 2021
- 2Self-Supervised Policy Adaptation during Deployment58 citations · 2020
- 3
- 4
- 5
- 6Visual Reinforcement Learning With Self-Supervised 3D Representations25 citations · 2023
- 7Generalization in Reinforcement Learning by Soft Data Augmentation16 citations · 2020
- 8
- 9Self-Supervised Policy Adaptation during Deployment8 citations · 2021
- 10MoDem-V2: Visuo-Motor World Models for Real-World Robot Manipulation5 citations · 2024