Name
Abstract | ||
|---|---|---|
The ability to plan into the future while utilizing only raw high-dimensional observations, such as images, can provide autonomous agents with broad capabilities. Visual model-based reinforcement learning (RL) methods that plan future actions directly have shown impressive results on tasks that require only short-horizon reasoning, however, these methods struggle on temporally extended tasks. We argue that it is easier to solve long-horizon tasks by planning sequences of states rather than just actions, as the effects of actions greatly compound over time and are harder to optimize. To achieve this, we draw on the idea of collocation, which has shown good results on long-horizon tasks in optimal control literature, and adapt it to the image-based setting by utilizing learned latent state space models. The resulting latent collocation method (LatCo) optimizes trajectories of latent states, which improves over previously proposed shooting methods for visual model-based RL on tasks with sparse rewards and long-term goals. See the videos on the supplementary website https: //orybkin.github.io/latco/. |
Year | Venue | Keywords |
|---|---|---|
2021 | INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 139 | Reinforcement learning,Intelligent agent,Collocation,Probabilistic logic,Optimal control,Visual control,Latent variable,Machine learning,Computer science,Artificial intelligence,Planning algorithms |
DocType | Volume | ISSN |
Conference | 139 | 2640-3498 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
6 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Oleh Rybkin | 1 | 0 | 4.39 |
| Chuning Zhu | 2 | 0 | 0.34 |
| Anusha Nagabandi | 3 | 24 | 4.41 |
| Konstantinos Daniilidis | 4 | 3122 | 255.45 |
| Igor Mordatch | 5 | 780 | 35.58 |
| Sergey Levine | 6 | 3377 | 182.21 |