Trial without Error: Towards Safe Reinforcement Learning via Human Intervention.
William SaundersGirish SastryAndreas StuhlmüllerOwain EvansPublished in: CoRR (2017)
Keyphrases
- reinforcement learning
- error rate
- machine learning
- learning algorithm
- function approximation
- website
- error bounds
- linear complexity
- state space
- reinforcement learning algorithms
- temporal difference
- temporal difference learning
- data sets
- partially observable
- model free
- learning problems
- markov decision processes
- learning process
- multi agent
- training data