Corpus ID: 220381033

Provably Safe PAC-MDP Exploration Using Analogies

@inproceedings{Roderick2021ProvablySP,
  title={Provably Safe PAC-MDP Exploration Using Analogies},
  author={Melrose Roderick and Vaishnavh Nagarajan and J. Z. Kolter},
  booktitle={AISTATS},
  year={2021}
}
A key challenge in applying reinforcement learning to safety-critical domains is understanding how to balance exploration (needed to attain good performance on the task) with safety (needed to avoid catastrophic failure). Although a growing line of work in reinforcement learning has investigated this area of "safe exploration," most existing techniques either 1) do not guarantee safety during the actual exploration process; and/or 2) limit the problem to a priori known and/or deterministic… Expand

Figures from this paper

AlwaysSafe: Reinforcement Learning without Safety Constraint Violations during Training

References

SHOWING 1-10 OF 44 REFERENCES
Safe Exploration and Optimization of Constrained MDPs Using Gaussian Processes
Approximate Exploration through State Abstraction
Safe Exploration in Finite Markov Decision Processes with Gaussian Processes
Safe Exploration in Markov Decision Processes
Safe exploration for reinforcement learning
Reachability-based safe learning with Gaussian processes
Safe Model-based Reinforcement Learning with Stability Guarantees
Model-based reinforcement learning with nearly tight exploration complexity bounds
Constrained Policy Optimization
Efficient reinforcement learning
...
1
2
3
4
5
...