提出MAGIC方法，精准估算删减训练数据对模型预测的影响。
MAGIC: Near-Optimal Data Attribution for Deep Learning
- 结合元微分与经典方法，实现近最优的数据影响估计
- 在大规模非凸场景下，估计结果与真实值相关性显著提升
- 适合需要可解释性与数据溯源的深度学习应用
预测型数据归属的目标是估算增删特定训练样本对模型预测的影响。在凸设置下，这一目标可通过无穷小刀切法直接实现。但在大规模非凸场景中，现有方法表现不佳——其估计结果与真实值的相关性较弱。本文提出一种新方法MAGIC，融合经典方法与元微分最新进展，近乎最优地估计增删训练数据对模型预测的影响。
原文摘要 · Abstract (English)
The goal of predictive data attribution is to estimate how adding or removing a given set of training datapoints will affect model predictions. In convex settings, this goal is straightforward (i.e., via the infinitesimal jackknife). In large-scale (non-convex) settings, however, existing methods are far less successful -- current methods' estimates often only weakly correlate with ground truth. In this work, we present a new data attribution method (MAGIC) that combines classical methods and recent advances in metadifferentiation to (nearly) optimally estimate the effect of adding or removing training data on model predictions.
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。