Alleviating limit cycling in training GANs with an optimization technique
作者机构:National Center for Applied Mathematics in ChongqingChongqing Normal UniversityChongqing401331China School of Mathematical SciencesUniversity of Electronic Science and Technology of ChinaChengdu611731China
出 版 物:《Science China Mathematics》 (中国科学(数学)(英文版))
年 卷 期:2024年第67卷第6期
页 面:1287-1316页
核心收录:
学科分类:12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 081104[工学-模式识别与智能系统] 08[工学] 0835[工学-软件工程] 0701[理学-数学] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术(可授工学、理学学位)]
基 金:supported by the Major Program of National Natural Science Foundation of China(Grant Nos.11991020 and 11991024) the Team Project of Innovation Leading Talent in Chongqing(Grant No.CQYC20210309536) NSFC-RGC(Hong Kong)Joint Research Program(Grant No.12261160365) the Scientific and Technological Research Program of Chongqing Municipal Education Commission(Grant No.KJQN202300528)
主 题:GANs general bilinear game predictive centripetal acceleration algorithm lower and upper complexity bounds PCAA-Adam
摘 要:In this paper,we undertake further investigation to alleviate the issue of limit cycling behavior in training generative adversarial networks(GANs)through the proposed predictive centripetal acceleration algorithm(PCAA).Specifically,we first derive the upper and lower complexity bounds of PCAA for a general bilinear game,with the last-iterate convergence rate notably improving upon previous ***,we combine PCAA with the adaptive moment estimation algorithm(Adam)to propose PCAA-Adam,for practical training of GANs to enhance their generalization ***,we validate the effectiveness of the proposed algorithm through experiments conducted on bilinear games,multivariate Gaussian distributions,and the CelebA dataset,respectively.