检索结果-南通市图书馆

TibetanGoTinyNet:a lightweight U-Net style network for zero learning of Tibetan Go

Frontiers of Information Technology & Electronic Engineering 2024年第7期25卷 924-937页

作者： Xiali LI Yanyin ZHANG Licheng WU Yandong CHEN Junzhi YU Key Laboratory of Ethnic Language Intelligent Analysis and Security Governance Ministry of EducationMinzu University of ChinaBeijing 100081China School of Information Engineering Minzu University of ChinaBeijing 100081China Department of Advanced Manufacturing and Robotics College of EngineeringPeking UniversityBeijing 100871China

The game of Tibetan Go faces the scarcity of expert knowledge and research ***,we study the zero learning model of Tibetan Go under limited computing power resources and propose a novel scaleinvariant U-Net style two-headed output lightweight network *** lightweight convolutional neural networks and capsule structure are applied to the encoder and decoder of TibetanGoTinyNet to reduce computational burden and achieve better feature extraction *** autonomous self-attention mechanisms are integrated into TibetanGoTinyNet to capture the Tibetan Go board’s spatial and global information and select important *** training data are generated entirely from self-play *** achieves 62%–78%winning rate against other four U-Net style models including Res-UNet,Res-UNet Attention,Ghost-UNet,and Ghost *** also achieves 75%winning rate in the ablation experiments on the attention mechanism with embedded positional *** model saves about 33%of the training time with 45%–50%winning rate for different monte–carlo tree search(MCTS)simulation counts when migrated from 9×9 to 11×11 *** for our model is available at https://***/paulzyy/TibetanGoTinyNet.

关键词： Zero learning Tibetan Go U-Net Self-attention mechanism Capsule network monte-carlo tree search

来源：

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

Rich-text document styling restoration via reinforcement learning

引用

Frontiers of Computer Science 2021年第4期15卷 93-103页

作者： Hongwei LI Yingpeng HU Yixuan CAO Ganbin ZHOU Ping LUO Key Lab of Intelligent Information Processing of Chinese Academy of Sciences(CAS) Institute of Computing TechnologyCASBeijing 100190China University of Chinese Academy of Sciences Beijing 100049China search Product Center WeChat Search Application DepartmentTencentBeijing 100080China

Richly formatted documents,such as financial disclosures,scientific articles,government regulations,widely exist on ***,since most of these documents are only for public reading,the styling information inside them is usually missing,making them improper or even burdensome to be displayed and edited in different formats and *** this study we formulate the task of document styling restoration as an optimization problem,which aims to identify the styling settings on the document elements,e.g.,lines,table cells,text,so that rendering with the output styling settings results in a document,where each element inside it holds the(closely)exact position with the one in the original *** that each styling setting is a decision,this problem can be transformed as a multi-step decision-making task over all the document elements,and then be solved by reinforcement ***,monte-carlo tree search(MCTS)is leveraged to explore the different styling settings,and the policy function is learnt under the supervision of the delayed *** a case study,we restore the styling information inside tables,where structural and functional data in the documents are usually *** shows that,our best reinforcement method successfully restores the stylings in 87.65%of the tables,with 25.75%absolute improvement over the *** also discuss the tradeoff between the inference time and restoration success rate,and argue that although the reinforcement methods cannot be used in real-time scenarios,it is suitable for the offline tasks with high-quality ***,this model has been applied in a PDF parser to support cross-format display.

关键词： styling restoration monte-carlo tree search reinforcement learning richly formatted documents tables

来源：

维普期刊数据库

同方期刊数据库评论

在线全文

学校读者我要写书评

暂无评论

A Simulation Study of Integrated Maneuver Trajectory Control

A Simulation Study of Integrated Maneuver Trajectory Control

引用

第32届中国控制与决策会议

作者： Pei Liu Hao Wang AVIC Xi’an Flight Automatic Control Research Institude

Aiming at the maneuver control problem that depends on the standard maneuver library or pilots experience,a method is proposed to improve the monte carlo tree search algorithm based on Receding Horizon *** establishing a search model to find the maneuver sequence,meanwhile updating the state of UAV by iteration until the maneuver target is *** simulation results show that maneuver missions can be completed with a high success rate in different conditions and for different ***,it is demonstrated that the algorithm is relatively robust with respect to parameters perturbation experiment.

关键词： Maneuver Flight Control Markov Decision Process monte-carlo tree search Receding Horizon Control

来源： cnki会议评论

在线全文

cnki会议

学校读者我要写书评

暂无评论

Elo Rating of Local Contextual Patterns

Elo Rating of Local Contextual Patterns

引用

2011 Chinese Control and Decision Conference（CCDC）

作者： Wenfeng Li,Zhiqing Liu Beijing University of Posts and Telecommunications,Beijing,China

Patterns are an essential method to improve the performance of computer Go *** learning algorithms for pattern extraction and evaluation are eternal topics for the study of computer Go *** paper presents a new machine learning algorithm to extract and evaluate contextual Go patterns with an Elo rating *** algorithm calculates the Elo rating of a contextual pattern by comparing it with patterns within a certain contextual area insetad of the whole *** only produce a more accurate evaluation of Go patterns,this algorithm also reduces the time complexity *** results indicate that Go pattern evaluations produced by the new algorithm result in stronger performance when used in simulation of monte-carlo tree search.

关键词： Local Contextual Patterns Global Contextual Patterns Elo Rating System monte-carlo tree search

来源： cnki会议评论

在线全文

cnki会议

学校读者我要写书评

暂无评论

UCT Based search in Phantom Go 25th

UCT Based Search in Phantom Go

引用

第25届中国控制与决策会议

作者： Shiquan Shen Jie Xu Hongye Li Tan Zbu Jiao Wang Dalian Measuring and Control Technique Research Institute College of Information Science and Engineering Northeastern University

ISBN: (纸本)9781467355339

Phantom Go is a kind of imperfect information two-player *** combines the large searching space in Go and the uncertainty of imperfect information games. monte carlo methods and UCT algorithm were suggested applying in Phantom Go and got certain *** applying monte carlo methods in Phantom Go,guessing opponent stones is the first step. In this paper, we discussed some special situations in Phantom Go to infer some of opponent’s stones based on the current information and gave a method for guessing opponent *** we applied UCT search using previous simulation results in Phantom Go and compared it with monte carlo tree search.

关键词： Phantom Go Imperfect Information Games UCT monte-carlo tree search

来源： cnki会议评论

在线全文

cnki会议

学校读者我要写书评

暂无评论

欢迎您,

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

在线全文

在线全文

在线全文

在线全文

在线全文

请选择保存的检索档案：

请选择收藏分类：

通借通还

欢迎您,

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

在线全文

在线全文

在线全文

在线全文

在线全文

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：