Survey on Counterfactual Reasoning in Automated Planning
A new survey paper examines how counterfactual reasoning can enhance automated planning by allowing deviations from initial task parameters. Traditional planning assumes fully specified initial states, goals, and actions, which works for deterministic domains but lacks flexibility. The paper categorizes existing works based on what elements are changed, when reasoning is triggered, and why and how changes are made. It concludes with key findings and open research questions for future work. The paper is available on arXiv under Computer Science > Artificial Intelligence.
Key facts
- Automated planning traditionally requires fully specified initial states, goals, and actions.
- Real-world planning often needs flexibility to deviate from original parameters.
- The paper surveys existing works on counterfactual reasoning in automated planning.
- Works are categorized by what elements are changed, when reasoning is triggered, and why and how changes are made.
- The paper discusses key findings and outlines open research questions.
- The paper is available on arXiv.
- The submission history is included.
- The paper is categorized under Computer Science > Artificial Intelligence.
Entities
Institutions
- arXiv