Reducing cost of service is an important goal for resource discovery and interaction technologies. The shortcomings of transhipment-method and hibernation-method are to increase holistic cost of service and to slower ...Reducing cost of service is an important goal for resource discovery and interaction technologies. The shortcomings of transhipment-method and hibernation-method are to increase holistic cost of service and to slower resource discovery respectively. To overcome these shortcomings, a context-aware computing-based method is developed. This method, firstly, analyzes the courses of devices using resource discovery and interaction technologies to identify some types of context related to reducing cost of service, then, chooses effective methods such as stopping broadcast and hibernation to reduce cost of service according to information supplied by the context but not the transhipment-method’s simple hibernations. The results of experiments indicate that under the worst condition this method overcomes the shortcomings of transhipment-method, makes the “poor” devices hibernate longer than hibernation-method to reduce cost of service more effectively, and discovers resources faster than hibernation-method; under the best condition it is far better than hibernation-method in all aspects.展开更多
Frame-Semantic Parsing(FSP)aims to extract frame-semantic structures from text.The task usually involves three subtasks sequentially:Target Identification(TI),Frame Identification(FI),and Frame Semantic Role Labeling(...Frame-Semantic Parsing(FSP)aims to extract frame-semantic structures from text.The task usually involves three subtasks sequentially:Target Identification(TI),Frame Identification(FI),and Frame Semantic Role Labeling(FSRL).The three subtasks are closely related while most previous studies model them individually,encountering error propagation and running efficiency problems.Recently,an end-to-end graphbased model is proposed to jointly process three subtasks in one model.However,it still encounters three problems:insufficient semantic modeling between targets and arguments,span missing,and lacking knowledge incorporation of FrameNet.To address the mentioned problems,this paper presents an End-to-end FSP model with Table Encoder(EFSP-TE),which models FSP as two semantically dependent region classification problems and extracts frame-semantic structures from sentences in a one-step manner.Specifically,EFSP-TE incorporates lexical unit knowledge into context encoder via saliency embedding,and develops an effective table representation learning method based on Biaffine network and multi-layer ResNetstyle-CNNs(Convolutional Neural Networks),which can fully exploit word-to-word interactions and capture the information of various levels of semantic relations between targets and arguments.In addition,it adopts two separate region-based modules to obtain potential targets and arguments,followed by two interactive classification modules to predict the frames and roles for the potential targets and arguments.Experiments on two public benchmarks show that the proposed approach achieves state-of-the-art performance in end-to-end setting.展开更多
文摘Reducing cost of service is an important goal for resource discovery and interaction technologies. The shortcomings of transhipment-method and hibernation-method are to increase holistic cost of service and to slower resource discovery respectively. To overcome these shortcomings, a context-aware computing-based method is developed. This method, firstly, analyzes the courses of devices using resource discovery and interaction technologies to identify some types of context related to reducing cost of service, then, chooses effective methods such as stopping broadcast and hibernation to reduce cost of service according to information supplied by the context but not the transhipment-method’s simple hibernations. The results of experiments indicate that under the worst condition this method overcomes the shortcomings of transhipment-method, makes the “poor” devices hibernate longer than hibernation-method to reduce cost of service more effectively, and discovers resources faster than hibernation-method; under the best condition it is far better than hibernation-method in all aspects.
基金supported by the Science and Technology Cooperation and Exchange Special Project of Shanxi Province(No.202204041101016)the Basic Research Program of Shanxi Province(No.202203021211286)+2 种基金the Four“Batches”Innovation Project of Invigorating Medical through Science and Technology of Shanxi Province(No.2022XM01)the Research Fund of Guangxi Key Lab of Multi-source Information Mining Security(No.MIMS22-10)the Foundation of Shanxi Vocational University of Engineering Science and Technology(No.KJ202203).
文摘Frame-Semantic Parsing(FSP)aims to extract frame-semantic structures from text.The task usually involves three subtasks sequentially:Target Identification(TI),Frame Identification(FI),and Frame Semantic Role Labeling(FSRL).The three subtasks are closely related while most previous studies model them individually,encountering error propagation and running efficiency problems.Recently,an end-to-end graphbased model is proposed to jointly process three subtasks in one model.However,it still encounters three problems:insufficient semantic modeling between targets and arguments,span missing,and lacking knowledge incorporation of FrameNet.To address the mentioned problems,this paper presents an End-to-end FSP model with Table Encoder(EFSP-TE),which models FSP as two semantically dependent region classification problems and extracts frame-semantic structures from sentences in a one-step manner.Specifically,EFSP-TE incorporates lexical unit knowledge into context encoder via saliency embedding,and develops an effective table representation learning method based on Biaffine network and multi-layer ResNetstyle-CNNs(Convolutional Neural Networks),which can fully exploit word-to-word interactions and capture the information of various levels of semantic relations between targets and arguments.In addition,it adopts two separate region-based modules to obtain potential targets and arguments,followed by two interactive classification modules to predict the frames and roles for the potential targets and arguments.Experiments on two public benchmarks show that the proposed approach achieves state-of-the-art performance in end-to-end setting.
