Corpus ID: 6398388

Exploration for Multi-task Reinforcement Learning with Deep Generative Models

@article{Bangaru2016ExplorationFM,
  title={Exploration for Multi-task Reinforcement Learning with Deep Generative Models},
  author={Sai Praveen Bangaru and J. S. Suhas and Balaraman Ravindran},
  journal={ArXiv},
  year={2016},
  volume={abs/1611.09894}
}
Exploration in multi-task reinforcement learning is critical in training agents to deduce the underlying MDP. Many of the existing exploration frameworks such as $E^3$, $R_{max}$, Thompson sampling assume a single stationary MDP and are not suitable for system identification in the multi-task setting. We present a novel method to facilitate exploration in multi-task reinforcement learning using deep generative models. We supplement our method with a low dimensional energy model to learn the… Expand
2 Citations
The Dreaming Variational Autoencoder for Reinforcement Learning Environments
  • 5
  • PDF
Multi-Task Deep Reinforcement Learning for Continuous Action Control
  • 36
  • PDF

References

SHOWING 1-9 OF 9 REFERENCES
Actor-Mimic: Deep Multitask and Transfer Reinforcement Learning
  • 344
  • PDF
Multi-task reinforcement learning: a hierarchical Bayesian approach
  • 230
  • PDF
Control of Memory, Active Perception, and Action in Minecraft
  • 197
  • PDF
Bayesian Multi-Task Reinforcement Learning
  • 86
  • PDF
Value Iteration Networks
  • 395
  • PDF
Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning
  • 2,841
  • PDF
An Uncertain Future: Forecasting from Static Images Using Variational Autoencoders
  • 364
  • PDF
Auto-Encoding Variational Bayes
  • 11,340
  • PDF
A Practical Guide to Training Restricted Boltzmann Machines
  • 2,414
  • PDF