Towards Return Parity in Markov Decision Processes

Keyphrases

• Markov Decision Process 100%
• Group Policy 75%
• State Space 50%
• Decomposition Theorem 50%
• Action Space 50%
• Group-wise 25%
• Between-group 25%
• Adverse Effects 25%
• High Stakes 25%
• Expected Time 25%
• Benchmark Dataset 25%
• Delay Effects 25%
• Temporal Domain 25%
• Reward Function 25%
• Demographic Groups 25%
• Fairness Concern 25%
• Lasting Impact 25%
• Discounted Reward 25%
• Recommender Systems 25%
• Machine Learning Models 25%
• Performance Disparities 25%
• Fairness Criteria 25%
• Process Sharing 25%
• Fairness Notion 25%
• Algorithmic Decision 25%
• Integral Probability Metrics 25%

Computer Science

• Markov Decision Process 100%
• And-States 50%
• Machine Learning 25%
• Recommender Systems 25%
• Demographic Group 25%