论文标题
预先训练的gan的数据修订
Data Redaction from Pre-trained GANs
论文作者
论文摘要
众所周知，大型预训练的生成模型偶尔会输出不良样本，这破坏了它们的可信度。减轻这种情况的常见方法是使用不同的数据或不同的正则化以不同的方式重新培训它们 - 它使用大量的计算资源，并不总是完全解决问题。 在这项工作中，我们采用了一种不同，更易于友好的方法，并研究了如何在训练后将模型置于模型之后，以使其“ redacts”或避免输出某些样本。我们表明，修订与数据删除根本不同，并且数据删除可能并不总是会导致修订。然后，我们考虑生成的对抗网络（GAN），并为数据修订提供了三种不同的算法，这些算法在描述了如何编辑样品的方式上不同。对现实世界图像数据集的广泛评估表明，我们的算法表现出色数据删除基准，并且能够编辑数据，同时以全面重新训练的成本的一小部分保留高生成质量。
Large pre-trained generative models are known to occasionally output undesirable samples, which undermines their trustworthiness. The common way to mitigate this is to re-train them differently from scratch using different data or different regularization -- which uses a lot of computational resources and does not always fully address the problem. In this work, we take a different, more compute-friendly approach and investigate how to post-edit a model after training so that it ''redacts'', or refrains from outputting certain kinds of samples. We show that redaction is a fundamentally different task from data deletion, and data deletion may not always lead to redaction. We then consider Generative Adversarial Networks (GANs), and provide three different algorithms for data redaction that differ on how the samples to be redacted are described. Extensive evaluations on real-world image datasets show that our algorithms out-perform data deletion baselines, and are capable of redacting data while retaining high generation quality at a fraction of the cost of full re-training.
