On the Limitations of Markovian Rewards to Express Multi-Objective, Risk-Sensitive, and Modal Tasks

Open in new window