arXiv:2605.26470v1 Announce Type: new 
Abstract: Generative posterior sampling using diffusion models has emerged as a dominant paradigm for solving inverse problems in imaging, which usually consists of three main components: data consistency (DC) guidance, classifier-free guidance (CFG) and stochasticity. While prior arts have focused on how to develop each or all components, less attention has given to how to schedule them, leading to heuristically fixed or partially adjusted suboptimal schedules. In this work, we argue that the interactions among all three components in terms of scheduling are crucial for significantly improved performance in solving inverse problems in imaging. Our analysis shows that aggressive CFG early in sampling conflict with DC guidance, while stochasticity brings the trajectory back to higher-probability regions. Based on these findings, we propose Triadic Dynamics Aware Posterior Sampling (TriPS), which reformulates posterior sampling as a time-varying control problem and optimizes schedules following a triadic trend of decreasing DC and stochasticity scales alongside increasing CFG scale. TriPS achieves this through two strategies: template-based search over functional priors for reliable baseline schedules, and Group Relative Policy Optimization (GRPO)-based reinforcement learning for more flexible temporal curves. Experiments demonstrate TriPS outperforms state-of-the-art baselines in data fidelity and perceptual realism.

تقدم دراسة جديدة نموذج (Triadic Dynamics Aware Posterior Sampling) المعروف باسم (TriPS)، الذي يعمل على تحسين جدولة توجيه اتساق البيانات وتوجيه خالٍ من المصنف والصدفة في عملية أخذ العينات الخلفية التوليدية باستخدام نماذج الانتشار لحل المشكلات العكسية في التصوير. تتناول هذه الطريقة القيود المفروضة على الجداول الثابتة أو المعدلة جزئيًا التي أعاقت الأداء في هذا المجال.

Un nuevo estudio presenta el Muestreo Posterior Consciente de Dinámicas Triádicas (TriPS), que optimiza la programación de la orientación de consistencia de datos, la orientación sin clasificador y la estocasticidad en el muestreo posterior generativo utilizando modelos de difusión para resolver problemas inversos en imágenes. Este enfoque aborda las limitaciones de los horarios fijos o parcialmente ajustados que han obstaculizado el rendimiento en esta área.

Une nouvelle étude présente le Triadic Dynamics Aware Posterior Sampling (TriPS), qui optimise la planification de l'orientation de la cohérence des données, de l'orientation sans classificateur et de la stochasticité dans l'échantillonnage postérieur génératif utilisant des modèles de diffusion pour résoudre des problèmes inverses en imagerie. Cette approche répond aux limitations des horaires fixes ou partiellement ajustés qui ont entravé les performances dans ce domaine.

A new study introduces Triadic Dynamics Aware Posterior Sampling (TriPS), which optimizes the scheduling of data consistency guidance, classifier-free guidance, and stochasticity in generative posterior sampling using diffusion models for solving inverse problems in imaging. This approach addresses the limitations of fixed or partially adjusted schedules that have hindered performance in this area.