通过学习弯曲插值路径，让生成模型走直线，提升采样速度。
Learning Straight Flows by Learning Curved Interpolants
- 用灵活插值替代线性插值，引导生成路径变直
- 无需模拟即可端到端优化，实现快速生成轨迹
- 适合追求高效生成的扩散模型研究者
流匹配模型通常使用线性插值定义前向加噪过程。这种设计结合噪声与目标分布的独立耦合，导致向量场常呈弯曲状，从而造成推断/生成过程缓慢。本文提出学习灵活（可能弯曲）的插值路径，以训练出直线型向量场，实现更快的生成。我们将其建模为多级优化问题，并提出一种高效的近似求解方法。该框架提供端到端、无需仿真即可优化的方案，可用来学习直线生成轨迹。
原文摘要 · Abstract (English)
Flow matching models typically use linear interpolants to define the forward/noise addition process. This, together with the independent coupling between noise and target distributions, yields a vector field which is often non-straight. Such curved fields lead to a slow inference/generation process. In this work, we propose to learn flexible (potentially curved) interpolants in order to learn straight vector fields to enable faster generation. We formulate this via a multi-level optimization problem and propose an efficient approximate procedure to solve it. Our framework provides an end-to-end and simulation-free optimization procedure, which can be leveraged to learn straight line generative trajectories.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。