HIGHLIGHTS
- who: T. and colleagues from the School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, Korea have published the research: Visual Pretraining via Contrastive Predictive Model for Pixel-Based Reinforcement Learning, in the Journal: Sensors 2022, 22, 6504. of /2022/
- what: As the authors show later in the experiments , pretraining the visual encoder with even in-domain data without considering dynamics is still underperformed. The study closest to ours is , which also attempted to learn the encoder detached from the RL objective. The soft policy improvement step then updates the stochastic policy . . .
If you want to have access to all the content you need to log in!
Thanks :)
If you don't have an account, you can create one here.