Self-Supervised Discovering of Interpretable Features for Reinforcement Learning.

@article{Shi2020SelfSupervisedDO,
  title={Self-Supervised Discovering of Interpretable Features for Reinforcement Learning.},
  author={Wenjie Shi and Gao Huang and Shiji Song and Zhuo-yuan Wang and Tingyu Lin and C. Wu},
  journal={IEEE transactions on pattern analysis and machine intelligence},
  year={2020},
  volume={PP}
}
  • Wenjie Shi, Gao Huang, +3 authors C. Wu
  • Published 2020
  • Computer Science, Medicine
  • IEEE transactions on pattern analysis and machine intelligence
Deep reinforcement learning (RL) has recently led to many breakthroughs on a range of complex control tasks. However, the decision-making process is generally not transparent. The lack of interpretability hinders the applicability in safety-critical scenarios. While several methods have attempted to interpret vision-based RL, most come without detailed explanation for the agent's behaviour. In this paper, we propose a self-supervised interpretable framework, which can discover causal features… Expand

References

SHOWING 1-10 OF 61 REFERENCES
Learn to Interpret Atari Agents
  • 12
  • PDF
Contrastive Explanations for Reinforcement Learning in terms of Expected Consequences
  • 31
  • PDF
Towards Interpretable Reinforcement Learning Using Attention Augmented Agents
  • 37
  • PDF
Visualizing and Understanding Atari Agents
  • 118
  • Highly Influential
  • PDF
Towards Better Interpretability in Deep Q-Networks
  • 12
  • PDF
Verifiable Reinforcement Learning via Policy Extraction
  • 82
  • PDF
Human-level control through deep reinforcement learning
  • 11,621
  • Highly Influential
  • PDF
Regularized Anderson Acceleration for Off-Policy Deep Reinforcement Learning
  • 5
  • PDF
Improving Robot Controller Transparency Through Autonomous Policy Explanation
  • Bradley Hayes, J. Shah
  • Computer Science
  • 2017 12th ACM/IEEE International Conference on Human-Robot Interaction (HRI
  • 2017
  • 134
  • PDF
...
1
2
3
4
5
...