Learning to Play Hard Exploration Games Using Graph-guided Self-navigation | |
Zhao EM(赵恩民)1,2; Yan RY(闫仁业)1,2; Li K(李凯)2; Li LJ(李丽娟)2; Xing JL(兴军亮)1,2 | |
2021-02 | |
会议日期 | 2021-02 |
会议地点 | 线上 |
DOI | 无 |
英文摘要 | Thisworkconsiderstheproblemofdeeprein-
forcementlearning(RL)withlongtimedependenciesands-
parserewards,asarefoundinmanyhardexplorationgames.
Agraph-basedrepresentationisproposedtoallowanagent
toperformself-navigationforenvironmentalexploration.The
graphrepresentationnotonlyeffectivelymodelstheenvironment
structure,butalsoefficientlytracestheagentstatechangesand
thecorrespondingactions.Byencouragingtheagenttoearna
newinfluence-basedcuriosityrewardfornewgameobservations,
thewholeexplorationtaskisdividedintosub-tasks,whichare
effectivelysolvedusingaunifieddeepRLmodel.Experimental
evaluationsonhardexplorationAtariGamesdemonstratethe
effectivenessoftheproposedmethod.Thesourcecodeand
learnedmodelswillbereleasedtofacilitatefurtherstudieson
thisproblem. |
学科主题 | 信息科学与系统科学 |
语种 | 英语 |
内容类型 | 会议论文 |
源URL | [http://ir.ia.ac.cn/handle/173211/52241] |
专题 | 融合创新中心_决策指挥与体系智能 |
通讯作者 | Xing JL(兴军亮) |
作者单位 | 1.SchoolofArtificialIntelligence,UniversityofChineseAcademyofSciences 2.Institute of Automation, Chinese Academy of Sciences |
推荐引用方式 GB/T 7714 | Zhao EM,Yan RY,Li K,et al. Learning to Play Hard Exploration Games Using Graph-guided Self-navigation[C]. 见:. 线上. 2021-02. |
个性服务 |
查看访问统计 |
相关权益政策 |
暂无数据 |
收藏/分享 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论