DropConnect is a recently introduced algorithm to prevent the co-adaptation of feature detectors. Compared to Dropout, DropConnect gains state-of-the-art results on several image recognition benchmarks. Motivated by the success of DropConnect, we extended this algorithm with the ability of sparse feature selection. In DropConnect algorithm, the dropping masks of weights are generated using Bernoulli gating variables that are independent of the weights and activations. We introduce a new strategy to generate masks depending on the outputs of previous layer. Using this method, neurons which are promising to produce sparser features will be assigned a bigger possibility to keep active in the forward and backward propagations. We then evaluate such sparsity constrained DropConnect on MNIST and CIFAR datasets in comparison with ordinary DropConnect and Dropout method. The results show that our new method improves the sparsity of features significantly, while not degrading the precision.

With the rapid development of social media, danmaku video provides a platform for users to communicate online. To some extent, danmaku video provides emotional timing information and an innovative method to analyze video data. In the age of big data, studying the characteristics of danmaku and its emotional tendencies can not only help us understand the psychological characteristics of users but also feedback the effective information of users to video platforms, which can help the platforms optimize related short video recommendations so that it can provide a more accurate solution for the selection of audiences during video production. However, danmaku is different from traditional comments. Current emotion classification methods are only suitable for two-dimensional classification which are not suitable for danmaku emotion analysis. Aiming at the problems such as the colloquialism, diversity, spelling errors, structural non-linearity informal language on the Internet, diversity of social topics, and context dependency of emotion analysis of the danmaku data, this paper proposes an improved emotion analysis model based on Bi-LSTM model to classify the further four-dimensional emotions of Pleasure, Anger, Sorrow and Joy. Furthermore, we add tags such as comment time and user name to the danmaku information. Experimental results show that the improved model has higher Accuracy, Recall, Precision, and F1-Score under the same conditions compared with the CNN and SVM. The classification effect of improved model is close to the SOTA. Experimental results also show that the improved model can be effectively applied to the analysis of irregular danmaku emotion.

Traditional cultural heritage, often damaged or limited to static displays, struggles to fully convey its profound significance. The rapid advancement of the Internet of Things (IoT) and Extended Reality (XR) technologies has introduced transformative solutions to the preservation, restoration, and interactive display of cultural heritage. IoT enables real-time data acquisition, multi-device connectivity, and dynamic interactions, while XR enhances cultural heritage engagement through immersive experiences and precise digital reconstructions. This paper provides a comprehensive review of the synergistic application of IoT and XR in cultural heritage, focusing on three key areas: 3D reconstruction and restoration, smart guidance, and immersive experiences. In 3D reconstruction and restoration, this review explores how IoT’s efficient data acquisition and transmission capabilities integrate with XR’s high-precision visualization techniques to improve accuracy and efficiency in heritage preservation. For Intelligent Guided Tour, this review analyzes how IoT-enabled device collaboration and real-time updates enhance XR-based personalized and interactive tour experiences. In immersive experiences, this review examines how IoT’s multi-source data sensing combined with XR’s multisensory interaction creates more engaging and participatory environments for users. The review also synthesizes existing user evaluation methodologies and identifies key challenges in the convergence of IoT and XR, including device compatibility, data transmission latency, cost, and privacy concerns. Future directions are proposed, such as advancing multimodal data processing, establishing open technical standards, and developing cost-effective solutions for broader adoption. This study highlights the immense potential of IoT and XR to revolutionize the preservation, understanding, and experience of cultural heritage, providing valuable insights for future research and applications.

Recently, to reduce the increasing energy consumption along with the explosive growth of smart terminals and network data in Internet of Things (IoT), green IoT is attracting more and more attention from both academia and industry. Towards green IoT, this paper devotes attention to Device-to-Device (D2D) communication and social network which are two essential components in green IoT and first reviews the latest developments about them. Further, this paper proposes a social interaction assisted resource sharing scheme for D2D communication, to improve the utilization of spectrum resources in green IoT. Specifically, the proposed scheme abstracts the D2D communication system into social layer and physical layer, by combining social network with D2D communication. Extensive simulations with real social interaction data are conducted and they show that the proposed scheme can achieve improvements in terms of the transmission success rate for green IoT.