STARC: A General Framework For Quantifying Differences Between Reward Functions.
Joar SkalseLucy FarnikSumeet Ramesh MotwaniErik JennerAdam GleaveAlessandro AbatePublished in: CoRR (2023)
Keyphrases
- reward function
- markov decision processes
- reinforcement learning
- inverse reinforcement learning
- policy search
- multiple agents
- transition probabilities
- statistically significant
- state variables
- markov decision process
- state space
- optimal policy
- reinforcement learning algorithms
- markov decision problems
- simple examples
- dynamical systems
- random walk
- linear programming
- data mining