Emotion classification in textual conversations focuses on classifying the emotion of each utterance from textual conversations. It is becoming one of the most important tasks for natural language processing in recent years. However, it is a challenging task for machines to conduct emotion classification in textual conversations because emotions rely heavily on textual context. To address the challenge, we propose a method to classify emotion in textual conversations, by integrating the advantages of deep learning and broad learning, namely DBL. It aims to provide a more effective solution to capture local contextual information (i.e., utterance-level) in an utterance, as well as global contextual information (i.e., speaker-level) in a conversation, based on Convolutional Neural Network (CNN), Bidirectional Long Short-Term Memory (Bi-LSTM), and broad learning. Extensive experiments have been conducted on three public textual conversation datasets, which show that the context in both utterance-level and speaker-level is consistently beneficial to the performance of emotion classification. In addition, the results show that our proposed method outperforms the baseline methods on most of the testing datasets in weighted-average F1.
{"title":"Deep Broad Learning for Emotion Classification in Textual Conversations","authors":"Sancheng Peng;Rong Zeng;Hongzhan Liu;Lihong Cao;Guojun Wang;Jianguo Xie","doi":"10.26599/TST.2023.9010021","DOIUrl":"https://doi.org/10.26599/TST.2023.9010021","url":null,"abstract":"Emotion classification in textual conversations focuses on classifying the emotion of each utterance from textual conversations. It is becoming one of the most important tasks for natural language processing in recent years. However, it is a challenging task for machines to conduct emotion classification in textual conversations because emotions rely heavily on textual context. To address the challenge, we propose a method to classify emotion in textual conversations, by integrating the advantages of deep learning and broad learning, namely DBL. It aims to provide a more effective solution to capture local contextual information (i.e., utterance-level) in an utterance, as well as global contextual information (i.e., speaker-level) in a conversation, based on Convolutional Neural Network (CNN), Bidirectional Long Short-Term Memory (Bi-LSTM), and broad learning. Extensive experiments have been conducted on three public textual conversation datasets, which show that the context in both utterance-level and speaker-level is consistently beneficial to the performance of emotion classification. In addition, the results show that our proposed method outperforms the baseline methods on most of the testing datasets in weighted-average F1.","PeriodicalId":60306,"journal":{"name":"Tsinghua Science and Technology","volume":"29 2","pages":"481-491"},"PeriodicalIF":6.6,"publicationDate":"2023-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/iel7/5971803/10258149/10258244.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"68027436","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-22DOI: 10.26599/TST.2023.9010004
Lei Xu;Junhai Zhai
Deep neural network (DNN) has strong representation learning ability, but it is vulnerable and easy to be fooled by adversarial examples. In order to handle the vulnerability of DNN, many methods have been proposed. The general idea of existing methods is to reduce the chance of DNN models being fooled by observing some designed adversarial examples, which are generated by adding perturbations to the original images. In this paper, we propose a novel adversarial example generation method, called DCVAE-adv. Different from the existing methods, DCVAE-adv constructs adversarial examples by mixing both explicit and implicit perturbations without using original images. Furthermore, the proposed method can be applied to both white box and black box attacks. In addition, in the inference stage, the adversarial examples can be generated without loading the original images into memory, which greatly reduces the memory overhead. We compared DCVAE-adv with three most advanced adversarial attack algorithms: FGSM, AdvGAN, and AdvGAN++. The experimental results demonstrate that DCVAE-adv is superior to these state-of-the-art methods in terms of attack success rate and transfer ability for targeted attack. Our code is available at https://github.com/xzforeverlove/DCVAE-adv.
{"title":"DCVAE-adv: A Universal Adversarial Example Generation Method for White and Black Box Attacks","authors":"Lei Xu;Junhai Zhai","doi":"10.26599/TST.2023.9010004","DOIUrl":"https://doi.org/10.26599/TST.2023.9010004","url":null,"abstract":"Deep neural network (DNN) has strong representation learning ability, but it is vulnerable and easy to be fooled by adversarial examples. In order to handle the vulnerability of DNN, many methods have been proposed. The general idea of existing methods is to reduce the chance of DNN models being fooled by observing some designed adversarial examples, which are generated by adding perturbations to the original images. In this paper, we propose a novel adversarial example generation method, called DCVAE-adv. Different from the existing methods, DCVAE-adv constructs adversarial examples by mixing both explicit and implicit perturbations without using original images. Furthermore, the proposed method can be applied to both white box and black box attacks. In addition, in the inference stage, the adversarial examples can be generated without loading the original images into memory, which greatly reduces the memory overhead. We compared DCVAE-adv with three most advanced adversarial attack algorithms: FGSM, AdvGAN, and AdvGAN++. The experimental results demonstrate that DCVAE-adv is superior to these state-of-the-art methods in terms of attack success rate and transfer ability for targeted attack. Our code is available at https://github.com/xzforeverlove/DCVAE-adv.","PeriodicalId":60306,"journal":{"name":"Tsinghua Science and Technology","volume":"29 2","pages":"430-446"},"PeriodicalIF":6.6,"publicationDate":"2023-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/iel7/5971803/10258149/10258153.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"68027597","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-22DOI: 10.26599/TST.2023.9010044
Lei Chen;Yuan Li;Yong Lei;Xingye Deng
Metapaths with specific complex semantics are critical to learning diverse semantic and structural information of heterogeneous networks (HNs) for most of the existing representation learning models. However, any metapaths consisting of multiple, simple metarelations must be driven by domain experts. These sensitive, expensive, and limited metapaths severely reduce the flexibility and scalability of the existing models. A metapath-free, scalable representation learning model, called Metarelation2vec, is proposed for HNs with biased joint learning of all metarelations in a bid to address this problem. Specifically, a metarelation-aware, biased walk strategy is first designed to obtain better training samples by using autogenerating cooperation probabilities for all metarelations rather than using expert-given metapaths. Thereafter, grouped nodes by the type, a common and shallow skip-gram model is used to separately learn structural proximity for each node type. Next, grouped links by the type, a novel and shallow model is used to separately learn the semantic proximity for each link type. Finally, supervised by the cooperation probabilities of all meta-words, the biased training samples are thrown into the shallow models to jointly learn the structural and semantic information in the HNs, ensuring the accuracy and scalability of the models. Extensive experimental results on three tasks and four open datasets demonstrate the advantages of our proposed model.
{"title":"Metarelation2vec: A Metapath-Free Scalable Representation Learning Model for Heterogeneous Networks","authors":"Lei Chen;Yuan Li;Yong Lei;Xingye Deng","doi":"10.26599/TST.2023.9010044","DOIUrl":"https://doi.org/10.26599/TST.2023.9010044","url":null,"abstract":"Metapaths with specific complex semantics are critical to learning diverse semantic and structural information of heterogeneous networks (HNs) for most of the existing representation learning models. However, any metapaths consisting of multiple, simple metarelations must be driven by domain experts. These sensitive, expensive, and limited metapaths severely reduce the flexibility and scalability of the existing models. A metapath-free, scalable representation learning model, called Metarelation2vec, is proposed for HNs with biased joint learning of all metarelations in a bid to address this problem. Specifically, a metarelation-aware, biased walk strategy is first designed to obtain better training samples by using autogenerating cooperation probabilities for all metarelations rather than using expert-given metapaths. Thereafter, grouped nodes by the type, a common and shallow skip-gram model is used to separately learn structural proximity for each node type. Next, grouped links by the type, a novel and shallow model is used to separately learn the semantic proximity for each link type. Finally, supervised by the cooperation probabilities of all meta-words, the biased training samples are thrown into the shallow models to jointly learn the structural and semantic information in the HNs, ensuring the accuracy and scalability of the models. Extensive experimental results on three tasks and four open datasets demonstrate the advantages of our proposed model.","PeriodicalId":60306,"journal":{"name":"Tsinghua Science and Technology","volume":"29 2","pages":"553-575"},"PeriodicalIF":6.6,"publicationDate":"2023-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/iel7/5971803/10258149/10258166.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"68027656","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In solving multi-objective vehicle routing problems with time windows (MOVRPTW), most existing algorithms focus on the optimization of a single problem formulation. However, little effort has been devoted to exploiting valuable knowledge from the alternate formulations of MOVRPTW for better optimization performance. Aiming at this insufficiency, this study proposes a decomposition-based multi-objective multiform evolutionary algorithm (MMFEA/D), which performs the evolutionary search on multiple alternate formulations of MOVRPTW simultaneously to complement each other. In particular, the main characteristics of MMFEA/D are three folds. First, a multiform construction (MFC) strategy is adopted to construct multiple alternate formulations, each of which is formulated by grouping several adjacent subproblems based on the decomposition of MOVRPTW. Second, a transfer reproduction (TFR) mechanism is designed to generate offspring for each formulation via transferring promising solutions from other formulations, making that the useful traits captured from different formulations can be shared and leveraged to guide the evolutionary search. Third, an adaptive local search (ALS) strategy is developed to invest search effort on different alternate formulations as per their usefulness for MOVRPTW, thus facilitating the efficient allocation of computational resources. Experimental studies have demonstrated the superior performance of MMFEA/D on the classical Solomon instances and the real-world instances.
{"title":"Solving Multi-Objective Vehicle Routing Problems with Time Windows: A Decomposition-Based Multiform Optimization Approach","authors":"Yiqiao Cai;Zifan Lin;Meiqin Cheng;Peizhong Liu;Ying Zhou","doi":"10.26599/TST.2023.9010048","DOIUrl":"https://doi.org/10.26599/TST.2023.9010048","url":null,"abstract":"In solving multi-objective vehicle routing problems with time windows (MOVRPTW), most existing algorithms focus on the optimization of a single problem formulation. However, little effort has been devoted to exploiting valuable knowledge from the alternate formulations of MOVRPTW for better optimization performance. Aiming at this insufficiency, this study proposes a decomposition-based multi-objective multiform evolutionary algorithm (MMFEA/D), which performs the evolutionary search on multiple alternate formulations of MOVRPTW simultaneously to complement each other. In particular, the main characteristics of MMFEA/D are three folds. First, a multiform construction (MFC) strategy is adopted to construct multiple alternate formulations, each of which is formulated by grouping several adjacent subproblems based on the decomposition of MOVRPTW. Second, a transfer reproduction (TFR) mechanism is designed to generate offspring for each formulation via transferring promising solutions from other formulations, making that the useful traits captured from different formulations can be shared and leveraged to guide the evolutionary search. Third, an adaptive local search (ALS) strategy is developed to invest search effort on different alternate formulations as per their usefulness for MOVRPTW, thus facilitating the efficient allocation of computational resources. Experimental studies have demonstrated the superior performance of MMFEA/D on the classical Solomon instances and the real-world instances.","PeriodicalId":60306,"journal":{"name":"Tsinghua Science and Technology","volume":"29 2","pages":"305-324"},"PeriodicalIF":6.6,"publicationDate":"2023-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/iel7/5971803/10258149/10258254.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"68028926","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-22DOI: 10.26599/TST.2023.9010034
Hongsong Chen;Xintong Han;Yiying Zhang
With the rapid development of information technologies, industrial Internet has become more open, and security issues have become more challenging. The endogenous security mechanism can achieve the autonomous immune mechanism without prior knowledge. However, endogenous security lacks a scientific and formal definition in industrial Internet. Therefore, firstly we give a formal definition of endogenous security in industrial Internet and propose a new industrial Internet endogenous security architecture with cost analysis. Secondly, the endogenous security innovation mechanism is clearly defined. Thirdly, an improved clone selection algorithm based on federated learning is proposed. Then, we analyze the threat model of the industrial Internet identity authentication scenario, and propose cross-domain authentication mechanism based on endogenous key and zero-knowledge proof. We conduct identity authentication experiments based on two types of blockchains and compare their experimental results. Based on the experimental analysis, Ethereum alliance blockchain can be used to provide the identity resolution services on the industrial Internet. Internet of Things Application (IOTA) public blockchain can be used for data aggregation analysis of Internet of Things (IoT) edge nodes. Finally, we propose three core challenges and solutions of endogenous security in industrial Internet and give future development directions.
{"title":"Endogenous Security Formal Definition, Innovation Mechanisms, and Experiment Research in Industrial Internet","authors":"Hongsong Chen;Xintong Han;Yiying Zhang","doi":"10.26599/TST.2023.9010034","DOIUrl":"https://doi.org/10.26599/TST.2023.9010034","url":null,"abstract":"With the rapid development of information technologies, industrial Internet has become more open, and security issues have become more challenging. The endogenous security mechanism can achieve the autonomous immune mechanism without prior knowledge. However, endogenous security lacks a scientific and formal definition in industrial Internet. Therefore, firstly we give a formal definition of endogenous security in industrial Internet and propose a new industrial Internet endogenous security architecture with cost analysis. Secondly, the endogenous security innovation mechanism is clearly defined. Thirdly, an improved clone selection algorithm based on federated learning is proposed. Then, we analyze the threat model of the industrial Internet identity authentication scenario, and propose cross-domain authentication mechanism based on endogenous key and zero-knowledge proof. We conduct identity authentication experiments based on two types of blockchains and compare their experimental results. Based on the experimental analysis, Ethereum alliance blockchain can be used to provide the identity resolution services on the industrial Internet. Internet of Things Application (IOTA) public blockchain can be used for data aggregation analysis of Internet of Things (IoT) edge nodes. Finally, we propose three core challenges and solutions of endogenous security in industrial Internet and give future development directions.","PeriodicalId":60306,"journal":{"name":"Tsinghua Science and Technology","volume":"29 2","pages":"492-505"},"PeriodicalIF":6.6,"publicationDate":"2023-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/iel7/5971803/10258149/10258245.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"68027655","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Solving Multi-Objective Vehicle Routing Problems with Time Windows: A Decomposition-Based Multiform Optimization Approach","authors":"","doi":"","DOIUrl":"https://doi.org/","url":null,"abstract":"","PeriodicalId":60306,"journal":{"name":"Tsinghua Science and Technology","volume":"29 2","pages":"1-11"},"PeriodicalIF":6.6,"publicationDate":"2023-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"68028925","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Exploring a Promising Region and Enhancing Decision Space Diversity for Multimodal Multi-objective Optimization","authors":"","doi":"","DOIUrl":"https://doi.org/","url":null,"abstract":"","PeriodicalId":60306,"journal":{"name":"Tsinghua Science and Technology","volume":"29 2","pages":"1-11"},"PeriodicalIF":6.6,"publicationDate":"2023-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"68027590","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-22DOI: 10.26599/TST.2023.9010015
Qihao Liu;Cuiyu Wang;Xinyu Li;Liang Gao
Considering both process planning and shop scheduling in manufacturing can fully utilize their complementarities, resulting in improved rationality of process routes and high-quality and efficient production. Hence, the study of Integrated Process Planning and Scheduling (IPPS) has become a hot topic in the current production field. However, when performing this integrated optimization, the uncertainty of processing time is a realistic key point that cannot be neglected. Thus, this paper investigates a Fuzzy IPPS (FIPPS) problem to minimize the maximum fuzzy completion time. Compared with the conventional IPPS problem, FIPPS considers the fuzzy process time in the uncertain production environment, which is more practical and realistic. However, it is difficult to solve the FIPPS problem due to the complicated fuzzy calculating rules. To solve this problem, this paper formulates a novel fuzzy mathematical model based on the process network graph and proposes a MultiSwarm Collaborative Optimization Algorithm (MSCOA) with an integrated encoding method to improve the optimization. Different swarms evolve in various directions and collaborate in a certain number of iterations. Moreover, the critical path searching method is introduced according to the triangular fuzzy number, allowing for the calculation of rules to enhance the local searching ability of MSCOA. The numerical experiments extended from the well-known Kim benchmark are conducted to test the performance of the proposed MSCOA. Compared with other competitive algorithms, the results obtained by MSCOA show significant advantages, thus proving its effectiveness in solving the FIPPS problem.
{"title":"Mathematical Modeling and a Multiswarm Collaborative Optimization Algorithm for Fuzzy Integrated Process Planning and Scheduling Problem","authors":"Qihao Liu;Cuiyu Wang;Xinyu Li;Liang Gao","doi":"10.26599/TST.2023.9010015","DOIUrl":"https://doi.org/10.26599/TST.2023.9010015","url":null,"abstract":"Considering both process planning and shop scheduling in manufacturing can fully utilize their complementarities, resulting in improved rationality of process routes and high-quality and efficient production. Hence, the study of Integrated Process Planning and Scheduling (IPPS) has become a hot topic in the current production field. However, when performing this integrated optimization, the uncertainty of processing time is a realistic key point that cannot be neglected. Thus, this paper investigates a Fuzzy IPPS (FIPPS) problem to minimize the maximum fuzzy completion time. Compared with the conventional IPPS problem, FIPPS considers the fuzzy process time in the uncertain production environment, which is more practical and realistic. However, it is difficult to solve the FIPPS problem due to the complicated fuzzy calculating rules. To solve this problem, this paper formulates a novel fuzzy mathematical model based on the process network graph and proposes a MultiSwarm Collaborative Optimization Algorithm (MSCOA) with an integrated encoding method to improve the optimization. Different swarms evolve in various directions and collaborate in a certain number of iterations. Moreover, the critical path searching method is introduced according to the triangular fuzzy number, allowing for the calculation of rules to enhance the local searching ability of MSCOA. The numerical experiments extended from the well-known Kim benchmark are conducted to test the performance of the proposed MSCOA. Compared with other competitive algorithms, the results obtained by MSCOA show significant advantages, thus proving its effectiveness in solving the FIPPS problem.","PeriodicalId":60306,"journal":{"name":"Tsinghua Science and Technology","volume":"29 2","pages":"285-304"},"PeriodicalIF":6.6,"publicationDate":"2023-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/iel7/5971803/10258149/10258253.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"68027593","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-22DOI: 10.26599/TST.2023.9010067
Canyun Dai;Xiaoyan Sun;Hejuan Hu;Yong Zhang;Dunwei Gong
The dispatch of integrated energy systems in coal mines (IES-CM) with mine-associated supplies is vital for efficient energy utilization and carbon emissions reduction. However, IES-CM dispatch is highly challenging due to its feature as multi-objective and strong multi-constraint. Existing constrained multi-objective evolutionary algorithms often fall into locally feasible domains with poorly distributed Pareto front, which greatly deteriorates dispatch performance. To tackle this problem, we transform the traditional dispatch model of IES-CM into two tasks: the main task with all constraints and the helper task with constraint adaptive. Then we propose a constraint adaptive multi-tasking differential evolution algorithm (CA-MTDE) to optimize these two tasks effectively. The helper task with constraint adaptive is developed to obtain infeasible solutions near the feasible domain. The purpose of this infeasible solution is to transfer guiding knowledge to help the main task move away from local search. Additionally, a dynamic dual-learning strategy using DE/current-to-rand/1 and DE/current-to-best/1 is developed to maintain task diversity and convergence. Finally, we comprehensively evaluate the performance of CA-MTDE by applying it to a coal mine in Shanxi Province, considering two IES-CM scenarios. Results demonstrate the feasibility of CA-MTDE and its ability to generate a Pareto front with exceptional convergence, diversity, and distribution.
{"title":"A Constraint Adaptive Multi-Tasking Differential Evolution Algorithm: Designed for Dispatch of Integrated Energy System in Coal Mine","authors":"Canyun Dai;Xiaoyan Sun;Hejuan Hu;Yong Zhang;Dunwei Gong","doi":"10.26599/TST.2023.9010067","DOIUrl":"https://doi.org/10.26599/TST.2023.9010067","url":null,"abstract":"The dispatch of integrated energy systems in coal mines (IES-CM) with mine-associated supplies is vital for efficient energy utilization and carbon emissions reduction. However, IES-CM dispatch is highly challenging due to its feature as multi-objective and strong multi-constraint. Existing constrained multi-objective evolutionary algorithms often fall into locally feasible domains with poorly distributed Pareto front, which greatly deteriorates dispatch performance. To tackle this problem, we transform the traditional dispatch model of IES-CM into two tasks: the main task with all constraints and the helper task with constraint adaptive. Then we propose a constraint adaptive multi-tasking differential evolution algorithm (CA-MTDE) to optimize these two tasks effectively. The helper task with constraint adaptive is developed to obtain infeasible solutions near the feasible domain. The purpose of this infeasible solution is to transfer guiding knowledge to help the main task move away from local search. Additionally, a dynamic dual-learning strategy using DE/current-to-rand/1 and DE/current-to-best/1 is developed to maintain task diversity and convergence. Finally, we comprehensively evaluate the performance of CA-MTDE by applying it to a coal mine in Shanxi Province, considering two IES-CM scenarios. Results demonstrate the feasibility of CA-MTDE and its ability to generate a Pareto front with exceptional convergence, diversity, and distribution.","PeriodicalId":60306,"journal":{"name":"Tsinghua Science and Technology","volume":"29 2","pages":"368-385"},"PeriodicalIF":6.6,"publicationDate":"2023-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/iel7/5971803/10258149/10258155.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"68027599","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-22DOI: 10.26599/TST.2023.9010043
Binbin Zhang;Li Wang
Recently, self-supervised learning has shown great potential in Graph Neural Networks (GNNs) through contrastive learning, which aims to learn discriminative features for each node without label information. The key to graph contrastive learning is data augmentation. The anchor node regards its augmented samples as positive samples, and the rest of the samples are regarded as negative samples, some of which may be positive samples. We call these mislabeled samples as “false negative” samples, which will seriously affect the final learning effect. Since such semantically similar samples are ubiquitous in the graph, the problem of false negative samples is very significant. To address this issue, the paper proposes a novel model, False negative sample Detection for Graph Contrastive Learning (FD4GCL), which uses attribute and structure-aware to detect false negative samples. Experimental results on seven datasets show that FD4GCL outperforms the state-of-the-art baselines and even exceeds several supervised methods.
{"title":"False Negative Sample Detection for Graph Contrastive Learning","authors":"Binbin Zhang;Li Wang","doi":"10.26599/TST.2023.9010043","DOIUrl":"https://doi.org/10.26599/TST.2023.9010043","url":null,"abstract":"Recently, self-supervised learning has shown great potential in Graph Neural Networks (GNNs) through contrastive learning, which aims to learn discriminative features for each node without label information. The key to graph contrastive learning is data augmentation. The anchor node regards its augmented samples as positive samples, and the rest of the samples are regarded as negative samples, some of which may be positive samples. We call these mislabeled samples as “false negative” samples, which will seriously affect the final learning effect. Since such semantically similar samples are ubiquitous in the graph, the problem of false negative samples is very significant. To address this issue, the paper proposes a novel model, False negative sample Detection for Graph Contrastive Learning (FD4GCL), which uses attribute and structure-aware to detect false negative samples. Experimental results on seven datasets show that FD4GCL outperforms the state-of-the-art baselines and even exceeds several supervised methods.","PeriodicalId":60306,"journal":{"name":"Tsinghua Science and Technology","volume":"29 2","pages":"529-542"},"PeriodicalIF":6.6,"publicationDate":"2023-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/iel7/5971803/10258149/10258249.pdf","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"68027659","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：