← 研踪 PaperTrailarXiv:2611.23038 · v1
When Anchors Shift the Target: Advantage-Gated Adaptation for Diffusion Reinforcement Learning
首次提交 2026-11-18 00:37
版本更新 2026-11-18 00:37
雷达观测 2026-11-18 16:38
摘要
Forward-process reinforcement learning adapts diffusion and flow-matching models by re-noising rewarded generations into regression targets. Because negative advantages induce concave updates with unbounded repulsive gradients, existing methods stabilize training using cached predictions from a moving policy copy, yet these anchors also alter what positive feedback teaches the model to fit. We formalize this conflict as anchor coupling: after separating frozen-reference regularizers, affine moving-anchor rules cannot eliminate negative curvature while preserving every positive regression target. We propose Advantage-Gated Adaptation (AGA) to decouple these roles. Positive samples retain standard target fitting, whose self-limiting update vanishes at the observed target independently of the moving anchor. Negative samples instead use the first-order tangent of the concave objective at the moving-policy prediction, preserving the local repulsion direction with a prediction-independent gradient. Within the quadratic-anchor family, these requirements uniquely determine the parameter-free gate . Because the moving copy provides local linearization rather than retention, AGA assigns long-horizon behavior preservation to a frozen pretrained reference regularized at its operating classifier-free guidance scale, interpreted through an implicit-reward lens. Across compositional generation, text rendering, and preference alignment, our experiments show that displacing positive targets degrades visual quality and heightens sensitivity to policy-copy schedules, whereas AGA consistently improves the reward--quality trade-off.
作者备注:12 pages, 4 figures, including supplementary material
观测到的版本
相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。
v1 · 2026-11-18 00:37 · 最新观测版本
When Anchors Shift the Target: Advantage-Gated Adaptation for Diffusion Reinforcement Learning
Forward-process reinforcement learning adapts diffusion and flow-matching models by re-noising rewarded generations into regression targets. Because negative advantages induce concave updates with unbounded repulsive gradients, existing methods stabilize training using cached predictions from a moving policy copy, yet these anchors also alter what positive feedback teaches the model to fit. We formalize this conflict as anchor coupling: after separating frozen-reference regularizers, affine moving-anchor rules cannot eliminate negative curvature while preserving every positive regression target. We propose Advantage-Gated Adaptation (AGA) to decouple these roles. Positive samples retain standard target fitting, whose self-limiting update vanishes at the observed target independently of the moving anchor. Negative samples instead use the first-order tangent of the concave objective at the moving-policy prediction, preserving the local repulsion direction with a prediction-independent gradient. Within the quadratic-anchor family, these requirements uniquely determine the parameter-free gate . Because the moving copy provides local linearization rather than retention, AGA assigns long-horizon behavior preservation to a frozen pretrained reference regularized at its operating classifier-free guidance scale, interpreted through an implicit-reward lens. Across compositional generation, text rendering, and preference alignment, our experiments show that displacing positive targets degrades visual quality and heightens sensitivity to policy-copy schedules, whereas AGA consistently improves the reward--quality trade-off.
查看此版本解读 →查看此版本原文 ↗