Recently, if you look at the trend of the Unmanned Combat Entities (UCE) on the world that are actually operational, a large number of personnel per one UCE has been operating the GCS. However, UCEs to perform the att...Recently, if you look at the trend of the Unmanned Combat Entities (UCE) on the world that are actually operational, a large number of personnel per one UCE has been operating the GCS. However, UCEs to perform the attack/reconnaissance mission are very expensive assets and require a considerable amount of time to train for UCE operations. Accordingly, the future battlefield environment has become important to develop multiple UCE ground control station. In this study, we developed a multiple UCE GCS that one operator can operate up to four UCEs. The software was built with a total of 6 displays using a Lockheed Martin Corporation’s prepar3D. Scenario of research takes into account the operation of the South Korea-type future multiple UCEs, to take advantage of the simulation system, in this paper, we propose for each of the concepts and technologies.展开更多
微博实体链接是把微博中给定的指称链接到知识库的过程,广泛应用于信息抽取、自动问答等自然语言处理任务(Natural language processing,NLP).由于微博内容简短,传统长文本实体链接的算法并不能很好地用于微博实体链接任务.以往研究大...微博实体链接是把微博中给定的指称链接到知识库的过程,广泛应用于信息抽取、自动问答等自然语言处理任务(Natural language processing,NLP).由于微博内容简短,传统长文本实体链接的算法并不能很好地用于微博实体链接任务.以往研究大都基于实体指称及其上下文构建模型进行消歧,难以识别具有相似词汇和句法特征的候选实体.本文充分利用指称和候选实体本身所含有的语义信息,提出在词向量层面对任务进行抽象建模,并设计一种基于词向量语义分类的微博实体链接方法.首先通过神经网络训练词向量模板,然后通过实体聚类获得类别标签作为特征,再通过多分类模型预测目标实体的主题类别来完成实体消歧.在NLPCC2014公开评测数据集上的实验结果表明,本文方法的准确率和召回率均高于此前已报道的最佳结果,特别是实体链接准确率有显著提升.展开更多
文摘Recently, if you look at the trend of the Unmanned Combat Entities (UCE) on the world that are actually operational, a large number of personnel per one UCE has been operating the GCS. However, UCEs to perform the attack/reconnaissance mission are very expensive assets and require a considerable amount of time to train for UCE operations. Accordingly, the future battlefield environment has become important to develop multiple UCE ground control station. In this study, we developed a multiple UCE GCS that one operator can operate up to four UCEs. The software was built with a total of 6 displays using a Lockheed Martin Corporation’s prepar3D. Scenario of research takes into account the operation of the South Korea-type future multiple UCEs, to take advantage of the simulation system, in this paper, we propose for each of the concepts and technologies.
文摘微博实体链接是把微博中给定的指称链接到知识库的过程,广泛应用于信息抽取、自动问答等自然语言处理任务(Natural language processing,NLP).由于微博内容简短,传统长文本实体链接的算法并不能很好地用于微博实体链接任务.以往研究大都基于实体指称及其上下文构建模型进行消歧,难以识别具有相似词汇和句法特征的候选实体.本文充分利用指称和候选实体本身所含有的语义信息,提出在词向量层面对任务进行抽象建模,并设计一种基于词向量语义分类的微博实体链接方法.首先通过神经网络训练词向量模板,然后通过实体聚类获得类别标签作为特征,再通过多分类模型预测目标实体的主题类别来完成实体消歧.在NLPCC2014公开评测数据集上的实验结果表明,本文方法的准确率和召回率均高于此前已报道的最佳结果,特别是实体链接准确率有显著提升.
