关于生成模型估计的可验证性准则
On distinguishability criteria for estimating generative models
摘要:
两个最近推出的标准,估计生成模型都是基于减少到二进制分类。噪声对比估计(NCE)是一种估计过程,其中生成模型被训练成能够区分数据样本和噪声样本。生成式对抗网络(GANs)是一对生成器网络和递归网络,生成器网络通过试图欺骗递归网络相信其样本是真实数据来学习生成样本。这两个估计过程使用相同的函数来驱动学习,这自然会引起关于它们如何相互关联的问题,以及该函数是否与极大似然估计(MLE)相关。NCE对应于训练属于{ \em}网络但使用固定生成器网络的内部数据模型。我们表明,一个变种的NCE,与一个动态的发电机网络,相当于极大似然估计。由于将学习的生成器与适当的动态选择的生成器配对恢复了MLE,因此可以预期相反的情况适用于将学习的生成器与特定的生成器配对。然而,我们发现,恢复MLE的学习发电机需要离开的可扩展性游戏。具体而言:(i)如果允许对NCE使用非平稳噪声分布,则可以使NCE的期望梯度与MLE的期望梯度相匹配,(ii)没有任何选择可以使GAN生成器的期望梯度与MLE的期望梯度相匹配,以及(iii)现有理论不能保证GANs在非凸情况下收敛。这表明GAN研究的关键下一步是确定GANs是否收敛,如果不收敛,则修改其训练算法以强制收敛。
英文摘要:
Two recently introduced criteria for estimation of generative models are both based on a reduction to binary classification. Noise-contrastive estimation (NCE) is an estimation procedure in which a generative model is trained to be able to distinguish data samples from noise samples. Generative adversarial networks (GANs) are pairs of generator and discriminator networks, with the generator network learning to generate samples by attempting to fool the discriminator network into believing its samples are real data. Both estimation procedures use the same function to drive learning, which naturally raises questions about how they are related to each other, as well as whether this function is related to maximum likelihood estimation (MLE). NCE corresponds to training an internal data model belonging to the {\em discriminator} network but using a fixed generator network. We show that a variant of NCE, with a dynamic generator network, is equivalent to maximum likelihood estimation. Since pairing a learned discriminator with an appropriate dynamically selected generator recovers MLE, one might expect the reverse to hold for pairing a learned generator with a certain discriminator. However, we show that recovering MLE for a learned generator requires departing from the distinguishability game. Specifically: (i) The expected gradient of the NCE discriminator can be made to match the expected gradient of MLE, if one is allowed to use a non-stationary noise distribution for NCE, (ii) No choice of discriminator network can make the expected gradient for the GAN generator match that of MLE, and (iii) The existing theory does not guarantee that GANs will converge in the non-convex case. This suggests that the key next step in GAN research is to determine whether GANs converge, and if not, to modify their training algorithm to force convergence.
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。