arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2401.09031cs.LG
收藏

扩散模型的数据归因:影响估计中由时间步引起的偏差

Data Attribution for Diffusion Models: Timestep-induced Bias in Influence Estimation

University of California, Los Angeles(加州大学洛杉矶分校)

Tong Xie, Haoyu Li, Andrew Bai, Cho-Jui Hsieh

AI总结:

针对扩散模型影响估计中由时间步引起的偏差问题,提出Diffusion-TracIn及其重归一化版本Diffusion-ReTrac,通过融合时间动态特性有效降低普遍影响力样本数量并提升检索针对性。

摘要:

数据归因方法将模型行为追溯至其训练数据集,为更好地理解“黑盒”神经网络提供了一种有效途径。尽管先前的研究已在多种场景下建立了模型输出与训练数据之间可量化的联系,但解释扩散模型输出与训练样本之间的关系仍未被充分探索。具体而言,扩散模型在一系列时间步上运行,而非先前场景中的瞬时输入输出关系,这对将现有框架直接扩展到扩散模型构成了重大挑战。值得注意的是,我们提出了Diffusion-TracIn,它融合了这种时间动态特性,并观察到样本的损失梯度范数高度依赖于时间步。这一趋势导致影响估计中出现显著偏差,对于在诱导大范数的时间步上训练的样本尤为明显,使其具有普遍的影响力。为减轻这种效应,我们引入了Diffusion-ReTrac作为一种重归一化适配方法,能够检索到更针对目标测试样本的训练样本,从而促进影响的局部化度量并提供更直观的可视化。我们通过各种评估指标和辅助任务证明了该方法的有效性,将具有普遍影响力的样本数量减少至原来的$\ rac{1}{3}$。

英文摘要:

Data attribution methods trace model behavior back to its training dataset, offering an effective approach to better understand ''black-box'' neural networks. While prior research has established quantifiable links between model output and training data in diverse settings, interpreting diffusion model outputs in relation to training samples remains underexplored. In particular, diffusion models operate over a sequence of timesteps instead of instantaneous input-output relationships in previous contexts, posing a significant challenge to extend existing frameworks to diffusion models directly. Notably, we present Diffusion-TracIn that incorporates this temporal dynamics and observe that samples' loss gradient norms are highly dependent on timestep. This trend leads to a prominent bias in influence estimation, and is particularly noticeable for samples trained on large-norm-inducing timesteps, causing them to be generally influential. To mitigate this effect, we introduce Diffusion-ReTrac as a re-normalized adaptation that enables the retrieval of training samples more targeted to the test sample of interest, facilitating a localized measurement of influence and considerably more intuitive visualization. We demonstrate the efficacy of our approach through various evaluation metrics and auxiliary tasks, reducing the amount of generally influential samples to $\frac{1}{3}$ of its original quantity.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑