Hands-On Deep Learning for Games
Thenumberofapplicationsofdeeplearningandneuralnetworkshasmultipliedinthelastcoupleofyears.Neuralnetshasenabledsignificantbreakthroughsineverythingfromcomputervision,voicegeneration,voicerecognitionandself-drivingcars.Gamedevelopmentisalsoakeyareawherethesetechniquesarebeingapplied.Thisbookwillgiveanindepthviewofthepotentialofdeeplearningandneuralnetworksingamedevelopment.Wewilltakealookatthefoundationsofmulti-layerperceptron’stousingconvolutionalandrecurrentnetworks.InapplicationsfromGANsthatcreatemusicortexturestoself-drivingcarsandchatbots.Thenweintroducedeepreinforcementlearningthroughthemulti-armedbanditproblemandotherOpenAIGymenvironments.AsweprogressthroughthebookwewillgaininsightsaboutDRLtechniquessuchasMotivatedReinforcementLearningwithCuriosityandCurriculumLearning.WealsotakeacloserlookatdeepreinforcementlearningandinparticulartheUnityML-Agentstoolkit.Bytheendofthebook,wewilllookathowtoapplyDRLandtheML-Agentstoolkittoenhance,testandautomateyourgamesorsimulations.Finally,wewillcoveryourpossiblenextstepsandpossibleareasforfuturelearning.
·8.2萬字