Online Learning in Markov Decision Processes with Adversarially Chosen Transition Probability Distributions

Mathematics

• Probability Distribution 100%
• Transition Probability 100%
• Markov Decision Process 100%
• Minimizes 50%
• Loss Function 50%
• Path Problem 50%
• Directed Acyclic Graph 25%
• Polynomial 25%
• Edge 25%
• Sample Path 25%
• Action Space 25%

Computer Science

• Electronic Learning 100%
• Markov Decision Process 100%
• Learning Algorithm 50%
• Efficient Algorithm 50%
• Shortest Path Problem 50%
• Process Problem 50%
• And-States 25%
• Transition Distribution 25%
• Directed Acyclic Graph 25%

Keyphrases

• Online Shortest Path 50%
• Transition Loss 25%
• Adversarial Loss 25%
• Cryptographic Scheme 25%
• Weighted Directed Acyclic Graphs 25%
• Action Space 25%
• Transition Distributions 25%