Related references
Note: Only part of the references are listed.
Article
Computer Science, Artificial Intelligence
Zhuo Lin Li et al.
Summary: Multivariate time series forecasting is a challenging task due to the complex dependencies between variables. Existing graph neural networks fail to handle this challenge as they either rely on predefined spatial graphs or fixed adjacency graphs. In this study, a novel framework called the static and dynamic graph learning network (SDGL) is proposed to model unknown association patterns. The SDGL approach infers both static and dynamic graph structures from the data to capture long-term and short-term patterns respectively. Experimental results show that our method achieves state-of-the-art performance on six benchmark datasets and effectively models dynamic spatio-temporal dependencies.
PATTERN RECOGNITION
(2023)
Article
Computer Science, Artificial Intelligence
Yizhang Wang et al.
Summary: Clustering based on Mutual K-nearest Neighbors (CMNN) is a classical method that groups data into clusters. However, it has limitations regarding the parameter k and misidentification of noise points. To address these issues, we propose an adaptive improved CMNN algorithm (AVCMNN) consisting of the improved CMNN algorithm (VCMNN) and the adaptive VCMNN algorithm (AVCMNN). The experimental results show that VCMNN and AVCMNN outperform other classical and state-of-the-art clustering algorithms.
PATTERN RECOGNITION
(2023)
Article
Computer Science, Artificial Intelligence
Di Jin et al.
Summary: Community detection is a fundamental task in network analysis, aiming to partition a network into sub-structures to reveal their latent functions. Traditional approaches utilize probabilistic graphical models, while new approaches use deep learning to convert networked data into low dimensional representation. There is a lack of understanding of the theoretical and methodological underpinnings of community detection, which is critical for the future development of network analysis.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2023)
Article
Computer Science, Artificial Intelligence
Jie Xu et al.
Summary: Incomplete multi-view clustering (IMVC) has gained increasing attention, however, existing methods still have issues regarding inaccurate imputed values and disregarding feature distribution discrepancy. To address this, we propose an imputation-free deep IMVC method that considers distribution alignment in feature learning. Our method utilizes autoencoders to learn view-specific features and implements adaptive feature projection to avoid imputing missing data. Experimental results demonstrate comparable or superior performance compared to state-of-the-art methods.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2023)
Article
Computer Science, Artificial Intelligence
Yonghua Zhu et al.
Summary: This paper proposes a GCN architecture combining interpretable feature learning and dynamic graph learning for personalized early Alzheimer's disease diagnosis, which outputs competitive diagnosis performance and provides interpretability.
INFORMATION FUSION
(2022)
Article
Computer Science, Artificial Intelligence
Tao Zhang et al.
Summary: This paper introduces a graph neural network model called Causal GraphSAGE (C-GraphSAGE) that incorporates causal inference into the sampling stage of GraphSAGE to improve classifier robustness. Experimental results demonstrate that C-GraphSAGE outperforms other graph neural network models in terms of classification performance when perturbation is present.
PATTERN RECOGNITION
(2022)
Article
Computer Science, Artificial Intelligence
Jiangzhang Gan et al.
Summary: This article proposes a novel multigraph fusion method to generate high-quality graphs and low-dimensional space for the GCN model, improving its robustness and representation learning capabilities.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Liang Peng et al.
Summary: Graph neural networks (GNNs) conduct feature learning by considering the local structure preservation of the data, but face issues with faulty and missing edges in the initial graph and the inability to generate predictions for unseen data. This study proposes a reverse GNN model to learn a high-quality graph from the intrinsic space of the original data points and introduces a new out-of-sample extension method.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Computer Science, Information Systems
Changan Yuan et al.
Summary: Subspace learning reduces the dimensions of high-dimensional data by projecting them into a low-dimensional subspace while preserving data similarity. Two strategies, reverse graph embedding and robust estimators, are designed to minimize the impact of redundancy and noise in the original space. Experiments show that this approach outperforms other state-of-the-art subspace learning methods across different evaluation metrics.
INFORMATION PROCESSING & MANAGEMENT
(2021)
Article
Computer Science, Artificial Intelligence
Wanxia Deng et al.
Summary: The Joint Clustering and Discriminative Feature Alignment (JCDFA) approach proposed in this paper aims to simultaneously mine discriminative features of target data and align cross-domain discriminative features to enhance performance in Unsupervised Domain Adaptation (UDA). The method integrates supervised classification of labeled source data and discriminative clustering of unlabeled target data, as well as optimizing supervised contrastive learning and conditional Maximum Mean Discrepancy (MMD) for feature alignment. Experimental results on real-world benchmarks demonstrate the superiority of JCDFA over state-of-the-art domain adaptation methods.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2021)
Article
Computer Science, Artificial Intelligence
Zhihui Li et al.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2018)