BROWSE
Detail
Conference Paper
Exploring Visual Reinforcement Learning for Sample-Efficient Robotic Manipulation
Cited 0 time in 
Share 
Share
- Authors
- Samyeul Noh, Seonghyun Kim, Ingook Jang
- Issue Date
- 2024-10
- Citation
- International Conference on Information and Communication Technology Convergence (ICTC) 2024, pp.207-210
- Publisher
- IEEE
- Language
- English
- Type
- Conference Paper
- Abstract
- Sample-efficient reinforcement learning (RL) from high-dimensional sensor readings holds significant potential for real physical robot applications. Recent developments in visual RL, which learns directly from image observations, have narrowed the gap between image-based and state-based training, especially in handling continuous actions. In this paper, we explore the potential of visual RL for achieving sample-efficient robotic manipulation. We evaluate four state-of-the-art visual RL algorithms-CURL, DrQ-v2, TACO, and DrM-on three robotic manipulation tasks from the DeepMind Control manipulation suite: 'reach-duplo,' 'push-box,' and 'lift-box.' Our findings reveal that while visual RL shows promising sample efficiency for robotic manipulation, it still lags behind state-based RL methods that learn from low-dimensional state vectors.
- KSP Keywords
- High-dimensional, Image-based, Recent developments, Reinforcement learning(RL), Robot applications, Robotic Manipulation, low-dimensional, state-based, state-of-The-Art