Event extraction is critical in various fields, including knowledge graph construction, public opinion monitoring, and situational awareness. Existing methods rely on supervised learning, which depends heavily on the ...
详细信息
ISBN:
(纸本)9798350379860;9798350379877
Event extraction is critical in various fields, including knowledge graph construction, public opinion monitoring, and situational awareness. Existing methods rely on supervised learning, which depends heavily on the scale and quality of the dataset, while the model's generalization ability remains weak. As a result, their application in real-world scenarios is significantly constrained. To address this issue, we propose a few-shot event extraction method that harnesses large language models' reading comprehension and event extraction capabilities. Our approach decomposes the task into multiple stages within a dialog-based framework. Experimental results show that our method outperforms most few-shot techniques, achieving a 4.1% performance improvement over ChatIE.
the widespread application of drone technology across multiple domains is significantly driving the development of low-altitude economy. Among these, the issue of end-to-end latency in drone communication networks is ...
详细信息
ISBN:
(纸本)9798350379860;9798350379877
the widespread application of drone technology across multiple domains is significantly driving the development of low-altitude economy. Among these, the issue of end-to-end latency in drone communication networks is particularly critical. We focus on the End-to-End(E2E) latency characteristics of UAV communication networks with emergent events, and use the concepts of arrival martingale and service martingale to estimate the latency boundary. Innovatively, the Doob-Meyer decomposition approach is introduced to decouple the stable factors and sudden events in the network, providing an accurate theoretical analysis tool for UAV communication network delays under stochastic network *** simulation experiments, it is proved that the proposed method can provide more accurate estimation of queue length and latency boundary while reducing computational complexity.
Knowledge graphs represent structured mappings of human knowledge playing a pivotal role in the iterative development and deepening of domain understandings. Methods for knowledge graph construction are continually ev...
详细信息
ISBN:
(纸本)9798350379860;9798350379877
Knowledge graphs represent structured mappings of human knowledge playing a pivotal role in the iterative development and deepening of domain understandings. Methods for knowledge graph construction are continually evolving. Advancements in artificial intelligence, particularly withthe advent of Generative Pre-trained Transformers (GPT), have paved the way for end-to-end automated methods for constructing knowledge graphs. this study provides a comprehensive evaluation of three conventional methodologies for constructing knowledge graphs: Top-Down, Bottom-Up and Joint approaches. We survey the associated technologies and present classic knowledge graph cases. We then elucidate the integration of large language models into the automation of knowledge graph construction, including their functions, benefits and limitations. Finally, we offer a prospective outlook on future research trajectories in this field.
Over the past decade, cloud computing has solidified its position as the primary platform for managing and processing extensive datasets within organizations. Despite its manifold advantages, integrating cloud computi...
详细信息
Graph Neural Networks (GNNs) have demonstrated their effectiveness in recommender systems. As a paradigm of self-supervised learning, the core of graph contrastive learning is to train the model by constructing positi...
详细信息
ISBN:
(纸本)9798350379860;9798350379877
Graph Neural Networks (GNNs) have demonstrated their effectiveness in recommender systems. As a paradigm of self-supervised learning, the core of graph contrastive learning is to train the model by constructing positive and negative samples, which is particularly important in datascarce environments and can effectively enhance the performance of recommender systems. the mainstream approaches in graph contrastive learning rely on user-item interaction graphs to generate contrastive views. However, most methods only consider various ways of creating these views, which, despite being straightforward, are prone to the effects of data noise and sparsity. In light of these issues, we propose a new graph contrastive learning framework called Comprehensive Graph Contrastive Learning for Recommendation (ComGCL). Our method considers boththe local and global aspects of the interaction graph, creating two contrastive views at different relational scales that account for the interaction graph's local and global features. We performed experiments across multiple datasets (***, Yelp, BeerAdvocate), and the results demonstrate that our model surpasses the existing state-of-the-art models in performance. Moreover, further experiments show that our model demonstrates greater robustness in the presence of data noise and sparsity.
User Experience (UX) refers to the individual experience of users when using software. An institution must analyze the user experience, especially for custom-built software. the goal of this study is to assess the UX ...
详细信息
this paper introduces an innovative method for low-complexity lossy source coding, according to duality between source and channel coding principles. Based on the Hamming code, we design hybrid Hamming code by utilizi...
详细信息
ISBN:
(纸本)9798350379860;9798350379877
this paper introduces an innovative method for low-complexity lossy source coding, according to duality between source and channel coding principles. Based on the Hamming code, we design hybrid Hamming code by utilizing the duality-based method. In the design process, it is shown that the syndrome of the hybrid Hamming code can be utilized as a complementary code to improve the reconstruction via the successive refinement link. this paper presents results of the simulations conducted to make a comparison in bit error rate (BER) performance between the hybrid Hamming code and the hybrid majority voting (HMV) code under different signal to noise ratio (SNR) in standalone links. Performance of hybrid Hamming code is also demonstrated in the case the coded sequences are transmitted over the refinement link.
In this paper, we investigate the problem of power allocation and user scheduling in the heterogeneous networks with base station sleep mode. Taking user communication quality and system energy consumption into accoun...
详细信息
ISBN:
(纸本)9798350379860;9798350379877
In this paper, we investigate the problem of power allocation and user scheduling in the heterogeneous networks with base station sleep mode. Taking user communication quality and system energy consumption into account, we aim to maximize energy efficiency of the system while satisfying both user load constraints and transmit power constraints of base stations. To decouple the original optimization problem into two subproblems, we put forward an iterative algorithm. We can use successive convex approximation and Lagrangian dual method to solve the first subproblem, and the second subproblem can be tackled using the discrete particle swarm optimization algorithm. Simulation results prove that the user scheduling and power allocation scheme proposed in this paper can effectively improve system energy efficiency while ensuring user communication quality, and reduce energy consumption of the system.
the proliferation of dense and heterogeneous networks (HetNets) due to the growing demand for mobile network traffic has a major impact on energy consumption in the telecommunications industry. Base stations (BS), as ...
详细信息
ISBN:
(纸本)9798350379860;9798350379877
the proliferation of dense and heterogeneous networks (HetNets) due to the growing demand for mobile network traffic has a major impact on energy consumption in the telecommunications industry. Base stations (BS), as primary energy consumers, account for approximately 70% of this energy use. this paper introduces a K-nearest neighbors (KNN)-based clustering strategy for managing the sleep control of small base stations (SBSs) in HetNets in an energy-efficient way. the KNN algorithm makes smart sleep/wake decisions by predicting traffic patterns using historical data. this approach effectively balances quality of service (QoS) requirements with energy savings, leading to significant reductions in power consumption. Simulation results demonstrate that this method outperforms existing solutions, making it a promising option for green communication in future wireless networks.
In the process of manned-unmanned coordinated combat, in order to give full play to the pilot's command advantages and the advantages of autonomous decision-making of manned aircraft and unmanned aircraft, it is n...
详细信息
ISBN:
(纸本)9798350379860;9798350379877
In the process of manned-unmanned coordinated combat, in order to give full play to the pilot's command advantages and the advantages of autonomous decision-making of manned aircraft and unmanned aircraft, it is necessary to reasonably allocate the decision-making functions among the three, so as to enable the system to give full play to the maximum combat effectiveness. this study takes the pilot-manned aircraft-unmanned aircraft in the manned-unmanned coordinated combat system as the research object, constructs the man-machine and machine-machine automation hierarchies, and combines the Uncertainty Linguistic Multi-Attribute Decision Making (ULMADM) method for the allocation of man-machine functions for typical combat tasks. the method takes the respective capability advantages of man-machine as inputs, and compared withthe traditional qualitative analysis, ULMADM can give the optimal automation level after comprehensively evaluating the combat scenario conditions related to the target functions.
暂无评论