Multivariate time series forecasting is a fundamental research problem in the Internet of Things (IoT), as it provides critical decision support and underpins the accuracy and reliability of downstream intelligent systems. Existing approaches commonly rely on sequence decomposition strategies that separate time series into trend, seasonal, and residual components. However, limited attention has been paid to how temporal information can be represented and integrated across different granularities. Inspired by the human reading process, in which information is progressively understood at the word, sentence, and paragraph levels, we propose LXformer, a novel forecasting framework that captures multiscale temporal representations. After segmenting multivariate time series into patches, LXformer integrates information at multiple granularities by modeling intra-patch features, inter-patch dependencies, and inter-sequence relationships to accomplish the forecasting task. Specifically, multiple one-dimensional convolutional branches are employed to extract fine-grained local patterns within each patch from diverse perspectives. In addition, agent attention is introduced to facilitate effective interactions across patches and channels, enabling the modeling of coarser-grained temporal dependencies. The combination of one-dimensional convolutions and linear-complexity attention mechanisms ensures that LXformer maintains overall linear computational complexity. Extensive experiments conducted on nine large-scale real-world datasets demonstrate that LXformer consistently achieves lower forecasting errors while delivering faster inference speed and reduced memory consumption. These advantages make LXformer particularly suitable for deployment on edge devices with limited computational resources but high accuracy requirements.

With the increasing development of ocean information technology, the multi-view fuzzy clustering is attracting increasing attention in pattern mining for massive multi-view ocean data of heterogeneous distributions, owing to its superior performance. However, the previous multi-view fuzzy clustering methods cannot fully consider informative topologies hidden in data distributions, which are crucial to recognize partitions of data. Moreover, they fail to capture invariant structures of multi-view ocean data in learning clustering-specific fusion representation. In addition, they do not take into consideration consistencies contained in the manifolds of data generation in mining soft patterns. To address those challenges, the deep multi-view generative fuzzy contrastive clustering (DMGFCC) is proposed within a Siamese architecture, which captures soft patterns of data via clustering-specific fusion representations of invariant structures in informative topologies. To be specific, a multi-view Siamese generative adversarial architecture is designed to capture the joint distribution of data as well as invariant structures, which is composed of the view-specific generator network providing pairwise implicit constraints, the view-specific discriminator network distilling knowledge of real data, and the view-specific cluster network capturing fuzzy patterns of fusion information. Furthermore, a generative adversarial dual contrastive clustering loss is devised, which consists of a generative adversarial loss fitting data distributions and a dual contrastive clustering loss learning soft patterns with consistencies of data manifolds. Finally, extensive experiments are conducted on four benchmark datasets, and the results demonstrate the competitive performance compared with the 11 representative methods.

Power systems are fundamental to state development, and measuring the security of the power information system is crucial. In existing power information systems, most security methods rely on encryption algorithms like RSA algorithm, which face growing threats from the increasing computing power of the Internet and quantum computers. With recognizing the urgent need for information security in power systems and the unconditional security offered by quantum communication, this paper explores the application of quantum security communication in power information systems. Specifically, through a detailed analysis of the procedure and security factors of the standard SSL protocol, this paper proposes a security enhancement method that is compatible with existing Internet protocol foundations. The proposed method enhances the source of random numbers by incorporating preset quantum random numbers, based on either a quantum random number generator or a quantum key distribution network. Implemented on the OPENSSL VPN evaluation platform, experiments show that the proposed security enhancement method can improve the security level of power information systems without significantly increasing system complexity or cost.