{"slug": "preference-elicitation-for-policy-optimization-and-application-to-aligning-heart", "title": "Preference Elicitation for Policy Optimization and Application to Aligning Heart Transplantation with Human Values", "summary": "A new preference elicitation algorithm for linear utilities outperforms prior techniques and, when applied to heart transplant allocation, yields policies near-optimal in alignment with human values, achieving a competitive ratio of 0.95 versus 0.54 for the status quo policy. The algorithm, developed by researchers (arXiv:2608.28620v1), learns a utility function from pairwise comparisons over outcomes, using cutting planes to warm-start and provably converging to the user's utility. In a user study, the method aggregated a community-aligned utility function to optimize transplant policies balancing post-transplant outcomes, waitlist mortality, geographic ease, and equity.", "body_md": "arXiv:2608.28620v1 Announce Type: new\nAbstract: Preference elicitation is essential for aligning AI systems with human values. Prior approaches (e.g., for organ allocation) often ask stakeholders to compare the decisions of an algorithm (e.g., patient A vs. patient B). Such a decision-level approach conflates the means with the ends. Instead, we elicit preferences directly over allocation outcomes to learn a utility function for policy optimization. We construct a novel preference elicitation algorithm for linear utilities that outperforms prior techniques in practice. Our algorithm has two phases. The first phase learns cutting planes through pairwise comparisons to rapidly shrink the space of possible attribute weights and warm-starts the second phase by eliminating dominated regions. The second phase then provably converges to the user's utility function. We apply our technique to heart transplant allocation where a policy must balance competing objectives such as post-transplant outcomes, waitlist mortality, geographic ease, and equity. Using our algorithm, we conduct a user study to learn and aggregate a community-aligned utility function, and use it to optimize heart transplant policies that are significantly better aligned with human values. Compared to the hindsight optimum, the status quo policy achieves a competitive ratio of just 0.54, while our method is near-optimal with a competitive ratio of 0.95.", "url": "https://wpnews.pro/news/preference-elicitation-for-policy-optimization-and-application-to-aligning-heart", "canonical_source": "https://arxiv.org/abs/2608.28620", "published_at": "2026-09-01 04:00:00+00:00", "updated_at": "2026-09-01 04:27:03.303720+00:00", "lang": "en", "topics": ["artificial-intelligence", "machine-learning", "ai-research", "ai-ethics"], "entities": ["arXiv"], "alternates": {"html": "https://wpnews.pro/news/preference-elicitation-for-policy-optimization-and-application-to-aligning-heart", "markdown": "https://wpnews.pro/news/preference-elicitation-for-policy-optimization-and-application-to-aligning-heart.md", "text": "https://wpnews.pro/news/preference-elicitation-for-policy-optimization-and-application-to-aligning-heart.txt", "jsonld": "https://wpnews.pro/news/preference-elicitation-for-policy-optimization-and-application-to-aligning-heart.jsonld"}}