扩散模型创造力源于得分平滑的插值效应

AI 摘要

Google Research 科学家郑道在 ICLR 2026 发表论文，证明扩散模型的创造力源于神经网络学习到的“平滑”得分函数，导致模型在训练数据点之间插值而非记忆。该研究结合正则化理论和去噪数学，解释了模型泛化的机制，并指出权重衰减等正则化方法会增强这种平滑效应。

正文节选

July 15, 2026 Zhengdao Chen, Research Scientist, Google Research We show that a diffusion model’s creativity (its ability to generate novel data, rather than just memorize its training set) is a mathematical consequence of neural networks learning a "smoothed" version of the score function, driving the model to interpolate between training data points along the hidden data manifold. Diffusion models are currently one of the most powerful types of tools for generative tasks that require complex a

发布时间：2026-07-16 02:06

抓取时间：2026-08-02 00:25

来源机构：Google Research