人工智能 机器学习Lecture-10.pptxVIP

人工智能核心能力培养—Lecture10

CNNArchitectures,ResNet,Inception,TransferLearning

NLPIntroduction

TheKernelAbilityForArtificialIntelligence高民权

review1.卷积神经网络的原理CNNPrinciplesFilterSize:WidthHeightFilterNumberChannelPaddingSizePoolingSizeFullyConnected2.卷积神经网络的计算过程与模型搭建方法CNNComputingProcedureandModelBuilding

today1.不同CNN模型的演化与对比DifferentCNNArchitectures2.迁移学习TransferLearning3.不同的优化器DifferentOptimizer4.剩余课程的介绍Introductiontheremaincourses5.自然语言处理的简介FirstIntroductionofNLP

Le-Net5YannLeCun,1998Conv5*5,s=1PoolingF=2,S=2Conv5*5,s=1P