协作式生成对抗网络  

Co-operative Generative Adversarial Nets

作　　者：张龙 赵杰煜[1] 叶绪伦 董伟 ZHANG Long;ZHAO Jie-Yu;YE Xu-Lun;DONG Wei(Faculty of Information Science and Engineering, Ningbo University, Ningbo 315211)

机构地区：[1]宁波大学信息科学与工程学院,宁波315211

出　　处：《自动化学报》2018年第5期804-810,共7页Acta Automatica Sinica

基　　金：国家自然科学基金(61571247);浙江省自然科学基金(LZ16F030001);浙江省国际合作项目(2013C24027)资助~~

摘　　要：生成对抗网络(Generative adversarial nets,GANs)将生成模型与判别模型进行了巧妙结合,采用无监督的训练方式,通过相互对抗共同提高,其在学术界掀起了一股新的机器学习热潮.GANs的学习目标是可以完整拟合任意真实样本的数据分布,然而在实际当中,真实样本分布的复杂程度难以预计,容易发生模式坍塌(Mode collapse)等问题,从而导致结果冗余,模型不收敛等.为提高无监督条件下的GANs生成能力,减少或消除模式坍塌,本文提出一种全新的协作式生成网络结构,通过构建多个生成模型,引入协作机制,使得生成模型在训练过程中能够相互学习,共同进步,从而提高模型对真实数据的拟合能力,进一步提高生成质量.通过在三组不同类型的数据集上进行实验,分析对比结果后发现新模型在二维图像生成方面,特别是人脸图片,有着显著的效果,协作机制不仅可以加快模型收敛速度,提高训练效率,还能消除损失函数噪声,在三维模型生成方面也产生了一定的影响.通过调整模型参数,模式坍塌问题也得到了遏制.本文还设计了一种动态学习方法,动态调节模型的学习速率,有效减少了过大或过小的梯度惩罚.Generative adversarial nets（GANs） combine the generative model with the discriminative model. With unsupervised training methods, the two types of models mutually improve through the adversarial process. It sets off a new machine learning boom in academia. The final goal of GANs learning is to fit any real-world data distribution. In practice, however, the real-world data distribution is difficult to estimate. The major problem is mode collapse, which may lead to redundancy and non-convergence. To improve the unsupervised generator and eliminate the risk of mode collapse,this paper proposes a novel co-operative network structure for GANs. Multiple generative models are constructed with a co-operative mechanism. It can help generative models to work together and learn from each other during training. In this way, the fitting ability of generators is largely enhanced, furthermore, the quality of generated data is eventually upgraded.Experiments are conducted on three different types of benchmark datasets. Results show that the new model significantly improves image generation, especially for human face pictures. Additionally, the co-operative mechanism can speed up the convergence, improve network s learning efficiency and deduct loss function noise. It also plays a certain role in 3D model generation and suppress the problem of mode collapse. In order to solve the inconsistency between generation model and discriminative model, a dynamic learning method is developed which can dynamically adjust learning frequency. It ultimately reduces unnecessary gradient penalties.

关 键 词：生成对抗网络 协作式 模式坍塌 生成模型 无监督学习 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]