arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2011.00754cs.LG
收藏

深度生成模型的泛化度量

Toward a Generalization Metric for Deep Generative Models

Deakin University(迪肯大学)

Hoang Thanh-Tung, Truyen Tran

摘要:

由于维数灾难,测量深度生成模型(DGMs)的泛化能力是困难的。DGMs的评估指标,如初始分数,Fréchet初始距离,精确召回和神经网络发散,试图使用多项式样本数估计生成的分布和目标分布之间的距离。这些指标是研究人员在设计新模型时的目标。尽管有这些说法,但仍然不清楚他们能在多大程度上衡量生成模型的泛化能力。在本文中,我们调查的能力,这些指标在衡量泛化能力。我们介绍了一个框架,比较评估指标的鲁棒性。我们表明,在这些指标更好的分数并不意味着更好的泛化。它们很容易被记住训练集的一小部分的生成器欺骗。我们建议修复NND度量,使其对生成的数据中的噪声更具鲁棒性。为了建立一个强大的度量泛化,我们建议将最小描述长度原则应用于评估DGMs的问题。我们开发了一种有效的方法来估计生成潜变量模型(GLVMs)的复杂性。实验结果表明,该度量可以有效地检测训练集记忆,区分不同泛化能力的GLVM。源代码可在https://github.com/htt210/GeneralizationMetricGAN上获得。

英文摘要:

Measuring the generalization capacity of Deep Generative Models (DGMs) is difficult because of the curse of dimensionality. Evaluation metrics for DGMs such as Inception Score, Fréchet Inception Distance, Precision-Recall, and Neural Net Divergence try to estimate the distance between the generated distribution and the target distribution using a polynomial number of samples. These metrics are the target of researchers when designing new models. Despite the claims, it is still unclear how well can they measure the generalization capacity of a generative model. In this paper, we investigate the capacity of these metrics in measuring the generalization capacity. We introduce a framework for comparing the robustness of evaluation metrics. We show that better scores in these metrics do not imply better generalization. They can be fooled easily by a generator that memorizes a small subset of the training set. We propose a fix to the NND metric to make it more robust to noise in the generated data. Toward building a robust metric for generalization, we propose to apply the Minimum Description Length principle to the problem of evaluating DGMs. We develop an efficient method for estimating the complexity of Generative Latent Variable Models (GLVMs). Experimental results show that our metric can effectively detect training set memorization and distinguish GLVMs of different generalization capacities. Source code is available at https://github.com/htt210/GeneralizationMetricGAN.

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。

↑