2022 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT)最新文献_第5页

A Context-Aware BERT Retrieval Framework Utilizing Abstractive Summarization 基于抽象摘要的上下文感知BERT检索框架

2022 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT) Pub Date : 2022-11-01 DOI: 10.1109/WI-IAT55865.2022.00142

Min Pan, Teng Li, Chenghao Yang, Shuting Zhou, Shaoxiong Feng, Youbin Fang, Xingyu Li

{"title":"A Context-Aware BERT Retrieval Framework Utilizing Abstractive Summarization","authors":"Min Pan, Teng Li, Chenghao Yang, Shuting Zhou, Shaoxiong Feng, Youbin Fang, Xingyu Li","doi":"10.1109/WI-IAT55865.2022.00142","DOIUrl":"https://doi.org/10.1109/WI-IAT55865.2022.00142","url":null,"abstract":"Recently, the multi-stage reranking framework based on pre-trained language model BERT can significantly improve the ranking performance on information retrieval tasks. However, most of these BERT-based reranking frameworks independently process query-chunk pairs and ignore cross-passages interaction. The context information around each candidate passage is extremely important for relevance judgement. Existing relevance aggregation methods obtain context information through statistical method and lost part of semantic information. Therefore, to capture this cross-passages interaction, this paper proposes a context-aware BERT ranking framework that utilizing abstractive summarization to enhance text semantics. By utilizing PEGASUS to summarize both sides of candidate passage accurately and then concatenate them as the input sequence, BERT could acquire more semantic information under the limitation of the input sequence’s length. The experimental results of two TREC data sets reveal the effectiveness of our proposed method in aggregating contextual semantic relevance.","PeriodicalId":345445,"journal":{"name":"2022 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT)","volume":"111 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125676905","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Visualization and Extraction of Important Structural Changes via Dynamic Hypergraph Embedding 基于动态超图嵌入的重要结构变化可视化与提取

2022 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT) Pub Date : 2022-11-01 DOI: 10.1109/WI-IAT55865.2022.00078

Shuta Ito, Takayasu Fushimi

{"title":"Visualization and Extraction of Important Structural Changes via Dynamic Hypergraph Embedding","authors":"Shuta Ito, Takayasu Fushimi","doi":"10.1109/WI-IAT55865.2022.00078","DOIUrl":"https://doi.org/10.1109/WI-IAT55865.2022.00078","url":null,"abstract":"Some real-world networks have structures that change dynamically over time. These changes include the addition or deletion of nodes and edges, and the rewiring of edges. Even when edge rewiring occurs, the degree of impact tends to vary depending on the location where it occurs and the nature of the node. In this study, we propose an embedding method that makes it easy to visually capture structural changes in dynamic hypergraphs. Furthermore, by quantifying the degree of influence of each hypernode, we attempt to extract influential structural changes that alter the location of many nodes in the network. Specifically, the positions of nodes are calculated by an embedding method that embeds hypernodes and hyperedges into the unit hypersphere based on their adjacencies, and the degree of influence on the nodes is calculated by the angle of the embedding vectors before and after the structural change occurs. We then propose a measure which is the average value of the influence degree of all nodes. Based on experimental evaluation using several synthetic datasets, we confirmed that our proposed measure quantifies the important structural changes as larger scores, conversely trivial changes as smaller ones.","PeriodicalId":345445,"journal":{"name":"2022 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125825560","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Algorithm for Cartography: Adding Out-domain Knowledge to the Level of Complexity Can Improve the Evolutionary Nature of Multilevel Genotype GIS 地图绘制算法:将域外知识添加到复杂层次可以提高多层次基因型GIS的进化特性

2022 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT) Pub Date : 2022-11-01 DOI: 10.1109/WI-IAT55865.2022.00146

Ziyang Weng, Xi Fang, Ziyu Zhang, Ren-yi Liu

{"title":"Algorithm for Cartography: Adding Out-domain Knowledge to the Level of Complexity Can Improve the Evolutionary Nature of Multilevel Genotype GIS","authors":"Ziyang Weng, Xi Fang, Ziyu Zhang, Ren-yi Liu","doi":"10.1109/WI-IAT55865.2022.00146","DOIUrl":"https://doi.org/10.1109/WI-IAT55865.2022.00146","url":null,"abstract":"By calculating the ecological efficacy in genotype-phenotype GIS, it can be concluded that evolution is the main attribute that explains the robustness and accessibility of genotype GIS. In this paper, we examine the definition of out-domain knowledge and how to enhance the level of complexity of genotype geographic system systems, through a multi-level computing model depending on road network growth characteristics in the description of the map-making process, realizing the mathematical logic expression, from data structures to regulatory networks to information clustering analysis. Our results suggest that historical archival information managed through spatiotemporal labels has many links to data in its seemingly unrelated socio-geographic information systems. Therefore, data showing high correlation and phenotypic abundance after location mapping are strongly cohesive, and common phenotypes are close to each other in genotype space. All of these properties are remarkable. Furthermore, evolutionary properties both increase with the number of genes in the genotype. The results show that increasing the complexity level of out-domain knowledge and increasing genome size can enhance both properties.","PeriodicalId":345445,"journal":{"name":"2022 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126008623","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Can higher-order structural features improve the performance of graph neural networks for graph classification? 高阶结构特征能提高图神经网络在图分类中的性能吗?

2022 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT) Pub Date : 2022-11-01 DOI: 10.1109/WI-IAT55865.2022.00130

Xin Chen, Miao Liu, Yue Peng, B. Shi

{"title":"Can higher-order structural features improve the performance of graph neural networks for graph classification?","authors":"Xin Chen, Miao Liu, Yue Peng, B. Shi","doi":"10.1109/WI-IAT55865.2022.00130","DOIUrl":"https://doi.org/10.1109/WI-IAT55865.2022.00130","url":null,"abstract":"Graph classification is a problem with applications in many different domains, which classifies a collection of graphs with categorical labels. One of the increasingly popular approaches to classify graphs is to use graph neural networks (GNNs), which capture the dependence of graph elements via message passing between the nodes. The key idea is to represent graphs in low-dimensional vectors by collectively aggregating node information guided by the graph structure. There are two types of information that can be used for graph classification. One is the textual features associated with each node in a graph, such as the keywords of a publication in a citation network. The other is the structural features that capture the higher-order dependencies between graph nodes. In this paper, we present a GNN-based graph classification framework that utilizes both textual and structural features, where the structural features of each node is calculated based on a set of small induced subgraphs (i.e., graphlets). We carry out experiments on several well-known graph-structured data sets, i.e., DD, MUTAG, NCI1, ENZYMES, and PROTEINS. By comparing with the state-of-the-art graph convolutional networks (GCNs), i.e., the spectral-based GCNs, the graph attention networks (GAT), and the transformer-based GCNs, we evaluate the effectiveness of involving graphlet-based structural features on the task of graph classification. The results also show that the transformer-based GCN, which integrates higher-order structural features as input, can significantly improve the accuracy of graph classification.","PeriodicalId":345445,"journal":{"name":"2022 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT)","volume":"91 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127049981","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

ConTED: Towards Content Trust for the Decentralized Web ted:走向去中心化网络的内容信任

2022 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT) Pub Date : 2022-11-01 DOI: 10.1109/WI-IAT55865.2022.00095

Valentin Siegert, Arved Kirchhoff, M. Gaedke

{"title":"ConTED: Towards Content Trust for the Decentralized Web","authors":"Valentin Siegert, Arved Kirchhoff, M. Gaedke","doi":"10.1109/WI-IAT55865.2022.00095","DOIUrl":"https://doi.org/10.1109/WI-IAT55865.2022.00095","url":null,"abstract":"The initiatives for redecentralization of the Web such as SoLiD aim to enhance users’ privacy by enforcing transparency about the data used by Web applications. However, it is a challenge for a Web application acquiring data from third-party sources to trust data originating from many or even hidden parties. A decentralized web application requires to evaluate trust and take trust-aware decisions autonomously without relying on a centralized infrastructure. While many related trust models consider direct or reputation-based trust for making trust-aware decisions, in decentralized web applications content and context factors (called content trust) become critical due to the arbitrary number of potential data providers and the contextual nature of trust. Besides, the dynamic nature of the de-centralized web necessitates trust-aware decisions that are made autonomously by the machine in a collaborative environment without further human intervention. To address these challenges, we present ConTED, a content trust evaluation framework for enabling decentralized Web applications to evaluate content trust autonomously. We also describe the architecture concept, which makes it feasible to integrate content trust models for decentralized Web applications. To demonstrate the feasibility, ConTED is integrated with aTLAS testbed, a web-based test bed to examine trust for a redecentralized web. Finally, we evaluate ConTED in terms of scalability and accuracy through a set of experiments.","PeriodicalId":345445,"journal":{"name":"2022 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT)","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124292293","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 1

A Survey of pet Action Recognition with action Recommendation based on HAR 基于HAR的宠物动作识别与动作推荐研究

2022 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT) Pub Date : 2022-11-01 DOI: 10.1109/wi-iat55865.2022.00125

Shunke Zhou

引用次数: 1

MTPL-G2T: Graph-to-Text Generation Task Based on Mixed Template Prompt Learning MTPL-G2T:基于混合模板提示学习的图到文本生成任务

2022 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT) Pub Date : 2022-11-01 DOI: 10.1109/WI-IAT55865.2022.00089

Jianhe Cen, Kun Zhang, Jingyuan Li, Shiqi Sun, Yuanzhuo Wang

引用次数: 0

Search on Asymmetric DCOPs by Strategic Agents 战略代理对非对称dcop的搜索

2022 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT) Pub Date : 2022-11-01 DOI: 10.1109/wi-iat55865.2022.00031

Yair Vaknin, A. Meisels

{"title":"Search on Asymmetric DCOPs by Strategic Agents","authors":"Yair Vaknin, A. Meisels","doi":"10.1109/wi-iat55865.2022.00031","DOIUrl":"https://doi.org/10.1109/wi-iat55865.2022.00031","url":null,"abstract":"Asymmetric Distributed Constraint Optimization Problems (ADCOPs) are a useful model for representing real-life problems of distributed nature. Constraining agents in ADCOPs have different gains (or costs) for the constraints that involve them. All former ADCOP search algorithms assume cooperation among the agents and do not capture the possibility of strategic behavior by the searching agents. The present paper extends a recent approach that uses side payments among constraining agents in ADCOP local search, and proposes an improved such algorithm for strategic agents. Enabling search for strategic agents is especially suitable for asymmetric DCOPs, where the agents gain differently from the constraints and would naturally pursue personal gains.The proposed method uses a specially designed mechanism that enforces truthful behavior for agents placing bids of side payments during search. This in turn guarantees that the (strategic) agents’ bids will form bids of maximal payoffs. The resulting search algorithm is an anytime algorithm that converges to stable solutions of higher social welfare that are local optima of the global social welfare, and computes the payments (contracts) that stabilize its outcome as a pure Nash equilibrium (PNE). The experimental evaluation shows that the payments charged by the mechanism in order to enforce truthful behavior are small compared to the total increase in the social welfare, and that the vast majority of the agents have improved personal gains when the algorithm terminates.","PeriodicalId":345445,"journal":{"name":"2022 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT)","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124710344","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Impact of Injecting Ground Truth Explanations on Relational Graph Convolutional Networks and their Explanation Methods for Link Prediction on Knowledge Graphs 注入基础真值解释对关系图卷积网络的影响及其对知识图链接预测的解释方法

2022 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT) Pub Date : 2022-11-01 DOI: 10.1109/WI-IAT55865.2022.00049

Nicholas F Halliwell, Fabien L. Gandon, F. Lécué

引用次数: 0

An Entire Space Multi-gate Mixture-of-Experts Model for Recommender Systems 推荐系统的全空间多门专家混合模型

2022 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT) Pub Date : 2022-11-01 DOI: 10.1109/WI-IAT55865.2022.00047

Zheng Ye, Jun Ge

{"title":"An Entire Space Multi-gate Mixture-of-Experts Model for Recommender Systems","authors":"Zheng Ye, Jun Ge","doi":"10.1109/WI-IAT55865.2022.00047","DOIUrl":"https://doi.org/10.1109/WI-IAT55865.2022.00047","url":null,"abstract":"With the development of e-commerce, both advertisers and platforms pay more and more attention to the effectiveness of ads recommendation. In recent years, deep learning approaches with a mulit-task learning framework have shown to be effective in such recommendation systems. One main goal of these systems is to estimate the post-click conversion rate(CVR) accurately. However, higher click-through rate(CTR) for a product does not always lead to higher conversion rate(CVR) due to many reasons (e.g. lower rating). In addition, the overall performance of the recommendation system may not be optimal, since the usage of multi-task models (the CTR and CVR tasks) is often sensitive to the relationships of the tasks. In this paper, we propose a deep neural model under the Mixture-of-Experts framework (MoE), call ES-MMOE, in which a sub-network is used to promote samples with high CVR. The model can also be trained with the entire space by taking advantage of the Entire Space Multi-task Model (ESMM) model. Extensive experiments on a large-scale dataset gathered from traffic logs of Taobao’s recommender system demonstrate that ES-MMOE outperforms a number of the state-of-the-art models, including ESMM, with a relatively large margin.","PeriodicalId":345445,"journal":{"name":"2022 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131140702","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0