Newest papers

CheckDistill: Adaptive Checklist Distillation for Efficient Distributional Reward Modeling in Image Editing

Yaqi Cai, Zuan Gao, Mengting Chen, Wei Li, Bo Rui Zheng · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.26740 · cited by 0 works (OpenAlex) · open-access PDF

Spanish

El aprendizaje de control de probabilidad para la edición de imágenes guiada por instrucciones requiere modelos de recompensa que proporcionen un feedback confiable, exacto y eficiente. La predicción escalar directa comprime las preferencias de edición multidimensional, mientras que los evaluadores generativos pueden exponer su juicio a través de un razonamiento explícito. Sin embargo, los criterios de puntuación amplios no especifican todos los requisitos de edición dependientes de la instancia y las restricciones de conservación, y la generación de trayectorias de razonamiento largas es costosa para la optimización en línea. Introducimos CheckDistill, un marco de modelado de recompensa puntual que internalizan las listas de control de instrucciones adaptativas. Descomponemos cada condición de edición en condiciones atómicas, requisitos verificables de forma independ

Machine translation (Meta M2M100); the English original follows.

Abstract

Reinforcement learning for instruction-guided image editing requires reward models that provide reliable, accurate, and efficient feedback. Direct scalar prediction compresses multidimensional editing preferences, whereas generative evaluators can expose their judgment through explicit reasoning. However, broad scoring criteria do not specify all instance-dependent editing and preservation constraints, and generating long reasoning trajectories is costly for online optimization. We introduce CheckDistill, a pointwise reward modeling framework that internalizes instruction-adaptive checklists. We decompose each editing condition into atomic, independently verifiable requirements and construct four-dimensional scoring labels together with structured verification evidence. After supervised score initialization, a privileged teacher conditions on checklists, item-level judgments, and dimension-level rationales to supervise the student's own scoring trajectories. A reliability-weighted on-policy self-distillation objective adjusts supervision at score tokens according to agreement between the teacher's expected scores and reference labels, while retaining candidate probability weights in a truncated token support. The deployed student directly predicts four score distributions without checklist inputs or explicit reasoning and aggregates their expectations into a scalar reward. CheckDistill achieves aggregate accuracies of 72.30%, 56.67% and 82.90% on MMRB2, MER-Bench and EditReward-Bench, respectively, exceeding the compared open-source baselines on all three metrics.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Computer science, Verifiable secret sharing, Machine learning, Artificial intelligence, Image editing, Security token

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange