3 papers
cs.RO2026
Collision Snapshot Guided Time-Reversed Safety-Critical Scenario Generation
Taehyung Kim, Jongeun Choi
The generation of safety-critical traffic scenarios is essential for training and evaluating autonomous vehicles. Prior approaches typically perturb the trajectories of existing ag…
cs.LG2026
To Go Far, Go Together: Diverse Preferences Induce a Curriculum for Reward Optimization
Taehyung Kim, Jongeun Choi
Learning a reward model from human feedback and optimizing a policy against it is one approach to aligning AI systems with individual users. From a fairness perspective, existing w…
cs.RO2026
Deployable Human Preference Alignment in Robotics: Learning Representative Rewards from Diverse Human Preferences
Taehyung Kim, Gwangmo Lee, Minjun Chang +2
Aligning robot policies with human preferences is essential for deployment to diverse end users. In per-user alignment approach, preference feedback is often sparse, so learning be…