Research map / Multi-agent systems and games / Games and equilibria
On the Expressivity of Markov Reward
David Abel, Will Dabney, Anna Harutyunyan, Mark K Ho, Michael Littman, Doina Precup, Satinder Singh
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Specify tasks as SOAPs, policy orderings, or trajectory orderings and require Markov reward to realize them | Finite CMPs · SOAPs, POs, and TOs | task expressibility | Mixed |
| Use range-SOAP instead of equal-SOAP realization | Markov reward design · SOAP realization | task expressibility | Improved |
| Design reward with a linear program for the specified task | Finite CMPs with SOAP, PO, or TO tasks · Reward design | computational complexity | Improved |
| Restrict reward functions to finitely many outputs | PO reward design · Finite-PO-RewardDecision | computational complexity | Hurt |
| Require one reward function to realize a task across multiple environments | A task realizable in each CMP separately · Sets of CMPs with shared state-action space | task expressibility | Hurt |
| Use a SOAP-designed reward instead of the regular goal-based reward | Q-learning in the grid world · Grid world | match to an acceptable policy | Improved |
| Increase the spread of acceptable policies | SOAP reward expressibility · Small CMPs | fraction of expressible SOAPs | Hurt |
Closest papers by meaning
- On the Expressivity of Objective-Specification Formalisms in Reinforcement Learning (ICLR 2024 · similarity 0.91)
- Explicable Reward Design for Reinforcement Learning Agents (NeurIPS 2021 · similarity 0.89)
- To the Max: Reinventing Reward in Reinforcement Learning (ICML 2024 · similarity 0.89)
- The Distributional Reward Critic Framework for Reinforcement Learning Under Perturbed Rewards (AAAI 2025 · similarity 0.88)
- Efficient Reinforcement Learning in Probabilistic Reward Machines (AAAI 2025 · similarity 0.88)
- Reinforcement Learning with Lookahead Information (NeurIPS 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
