arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1810.01876cs.LGstat.ML
收藏

深度生成模型中的虚假样本:bug还是特征?

Spurious samples in deep generative models: bug or feature?

CNRS/Université Paris-Saclay(法国国家科学研究中心/巴黎-萨克雷大学);PSL Research University(巴黎文理研究大学)

Balázs Kégl, Mehdi Cherti, Akın Kazakçı

摘要:

生成式建模文献中的传统观点是,模型可能生成的虚假样本是错误的,应该避免。然而,最近的研究表明,人们有兴趣研究甚至利用这些样本,而不是消除它们。在本文中,我们问这样的样本是否可以一起消除,而不牺牲生成分布的覆盖范围的问题。对于我们考虑的模型类,我们通过实验证明,这是不可能的,而不会失去模拟一些测试样本的能力。虽然我们的结果需要在更广泛的模型家族上得到证实,但这些初步发现提供了部分证据,证明虚假样本与学习数据集具有相同的结构特性,这反过来表明它们不仅仅是错误,而是深度生成网络的一个特征。

英文摘要:

Traditional wisdom in generative modeling literature is that spurious samples that a model can generate are errors and they should be avoided. Recent research, however, has shown interest in studying or even exploiting such samples instead of eliminating them. In this paper, we ask the question whether such samples can be eliminated all together without sacrificing coverage of the generating distribution. For the class of models we consider, we experimentally demonstrate that this is not possible without losing the ability to model some of the test samples. While our results need to be confirmed on a broader set of model families, these initial findings provide partial evidence that spurious samples share structural properties with the learned dataset, which, in turn, suggests they are not simply errors but a feature of deep generative nets.

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。

↑