ManifoldCache：基于约束流形缓存的免训练扩散加速
ManifoldCache: Training-Free Diffusion Acceleration via Constraint Manifold Caching
AI 导读
arXiv 论文提出 ManifoldCache，这是首个面向约束流形扩散模型（CMDMs）的免训练、免数据加速器。方法将条件得分正交分解为法向与切向分量，并证明噪声调度中点是安全缓存边界：边界前缓存误差有界，之后则必然导致部分轨迹出现模式混淆；更深的网络块可承受更大的认证缓存步长。该方案覆盖医学体积、分子构象、蛋白质骨架、晶体结构与多视角 3D 场景等八类模型，无需校准数据且零训练开销。
正文 · 原文
Abstract:Diffusion models for structured scientific generation must produce samples satisfying hard geometric constraints imposed by physics, chemistry, or biology, yet inference in these settings is prohibitively slow, demanding hundreds to thousands of neural-function evaluations per sample. We unify eight state-of-the-art models spanning medical volumetrics, molecular conformations, protein backbone design, crystal structure prediction, and multi-view 3D scenes under a single abstraction, Constraint-Manifold Diffusion Models (CMDMs), in which the target distribution is supported on a manifold defined by an externally specified constraint map. All existing acceleration families fail on this class: quantization exhausts memory on high-dimensional volumetric operators; pruning breaks constraint fidelity; fast ODE solvers allow trajectories to drift off the constraint manifold; and feature-caching heuristics are blind to constraint geometry, inducing mode confusion in the high-noise regime. We introduce ManifoldCache, the first training-free, data-free accelerator designed from first principles for CMDMs. The key insight is that the conditional score decomposes orthogonally into a normal component, which enforces constraint satisfaction, and a tangential component, which navigates within the manifold. Exploiting this structure, we prove that the noise-schedule midpoint is a sharp safe-caching boundary: caching before it incurs provably bounded error, while caching after it guarantees a strictly positive fraction of trajectories suffer mode confusion, a gap that persists up to the boundary. We further prove that deeper network blocks admit provably larger certified cache strides within the safe phase, as a consequence of the score decomposition propagating through block Jacobians. The resulting schedule requires no calibration data, along with zero training overhead.
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.04510 [cs.AI] |
| (or arXiv:2610.04510v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.04510 arXiv-issued DOI via DataCite |
Submission history
From: Prashant Pandey [view email]
[v1]
Sat, 3 Oct 2026 13:18:15 UTC (4,174 KB)
[v2]
Tue, 27 Oct 2026 18:26:59 UTC (3,167 KB)
来源：arXiv cs.AI · arxiv.org