Multi-modal knowledge graphs (MMKGs) enrich the semantic expression capabilities of traditional knowledge graphs by incorporating diverse modal information, showcasing immense potential in various knowledge reasoning tasks. However, existing MMKGs encounter numerous challenges in the link prediction task (i.e., knowledge graph completion reasoning), primarily due to the complexity and diversity of modal information and the imbalance in quality. These challenges make the efficient fusion and enhancement of multi-modal information difficult to achieve. Most existing methods adopt simple concatenation or weighted fusion of modal features, but such approaches fail to fully capture the deep semantic interactions between modalities and perform poorly when confronted with modal noise or missing information. To address these issues, this paper proposes a novel framework model-Adaptive Fusion and Modality Information Enhancement(AFME). This framework consists of two parts: the Modal Information Fusion module (MoIFu) and the Modal Information Enhancement module (MoIEn). By introducing a relationship-driven denoising mechanism and a dynamic weight allocation mechanism, the framework achieves efficient adaptive fusion of multi-modal information. It employs a generative adversarial network (GAN) structure to enable global guidance of structural modalities over feature modalities and adopts a multi-layer self-attention mechanism to optimize both intra- and inter-modal features. Finally, it jointly optimizes the losses of the triple prediction task and the adversarial generation task. Experimental results demonstrate that the AFME framework significantly improves multi-modal feature utilization and knowledge reasoning capabilities on multiple benchmark datasets, validating its efficiency and superiority in complex multi-modal scenarios.

As multilingual artificial intelligence systems proliferate, achieving robust cross-lingual understanding remains an open challenge. Recent works have made progress on visual question answering (VQA) models by pretraining on large English image-text datasets. However, there is a language gap as most models are English-centric. Existing attempts at multilingual VQA rely on machine translation or multilingual model pretraining, but cannot effectively transfer rich cross-modal knowledge from English models. In this work, we propose the cross-lingual multimodal knowledge transfer (CMKT) framework to efficiently extend English VQA models to non-English languages via knowledge distillation. Specifically, we introduce a code-mixed cross-lingual mask modeling (CCM) method to establish representations for new languages using small image-text data. We also design a multimodal knowledge distillation (MMKD) method to transfer modal understanding from English models by imitating their sequence processing. Experiments on the xGQA benchmark demonstrate that CMKT can effectively improve zero-shot learning and few-shot learning in non-English languages. Our method reduces the data and computation needed to train multilingual VQA models from scratch. The knowledge transfer paradigm enables non-English languages to inherit and generalize the intricate visual-semantic relationships learned from English. The results also show that the proposed method outperforms previous state-of-the-art methods in the zero-shot setting on the xGQA dataset.

Large Multimodal Models (LMMs) excel in English multimedia tasks but face challenges in adapting to other languages due to linguistic diversity, limited non-English multimodal data, and high training costs. Existing approaches rely on machine-translated multimodal corpora or multilingual large language models, yet they demand substantial resources and achieve only modest zero-shot cross-lingual transfer performance, as shown in the IGLUE benchmark. In this work, we propose SMSA, a Syntax-aware Multimodal Semantic Adaptation approach, which efficiently extends vision-language models (VLMs) to multiple languages via a lightweight adaptation module. Instead of learning from scratch, SMSA transfers multimodal knowledge from English-trained models using two key components: (1) a Syntax-aware Adapter (SAA), which restructures multilingual text representations to align better with English syntax, reducing cross-lingual misalignment; (2) a Multimodal Semantic Distillation (MSD) method, which enables the model to mimic English sequence processing and retain multimodal associations across languages. This allows efficient adaptation to new languages while preserving the original model's strong multimodal capabilities. We extend an MoE-based VLM to 8 languages using a small translation dataset. Evaluations on the IGLUE benchmark show that SMSA achieves strong zero-shot transfer, outperforming some multilingual LMMs and demonstrating its effectiveness in cross-lingual vision-language adaptation.

Mobile edge networks enhance cloud applications by offering reduced delay for mobile terminals. However, due to the cumulative increase of mobile terminals and applications, edge networks fail to accommodate all applications and are difficult to guarantee the end-to-end delay demand, which is an important metric of QoS. One potential solution to address this challenge is to place virtual clusters (VCs) in mobile micro-clouds (MMCs) to enhance the capabilities of edge networks and the cloud. Therefore, in this paper, we build the optimization model of VC placement in edges for maximizing the benefit function, with constraints of 1). the satisfaction for end-to-end delay demand of each application; 2). the resource capacity of each MMC no more than the whole demands of those VCs placed in. We use Stochastic Network Calculus (SNC) to build the delay model more accurately, considering stochastic arrivals of flows and stochastic services of service nodes. Moreover, we use Martingale Theory to enhance the solvability of the SNC-based delay model, achieving tighter delay bound reliability.

Knowledge discovery is an ongoing research endeavor aimed at uncovering valuable insights and patterns from large volumes of data in massive social systems (MSSs). Although recent advances in deep learning have made significant progress in knowledge discovery, the “data dimensionality reduction” problem still poses practical challenges. To address this, we have introduced a hierarchical information compression (IC) approach, which emphasizes the elimination of redundant and irrelevant features and the generation of high-quality knowledge representation, aiming to enhance the information density of the knowledge discovery process. Our approach consists of coarse-grained and fine-grained stages for data compression. In the coarse-grained stage, our method employs the key feature distiller based on the Siamese network to effectively identify a substantial number of irrelevant features and latent redundancies within coarse-grained data blocks. Moving on to the fine-grained stage, our model further compresses the internal features of the data, extracting the most crucial knowledge and facilitating data compression by cross-block learning. By implementing these two stages, the approach achieves both inter and innerblock IC while preserving essential knowledge. To validate the performance of our proposed model, we conducted several experiments using WikiSum, a large knowledge corpus based on English Wikipedia in MSSs. The experimental results demonstrate that our model achieved a 2.38% increase on recall-oriented understudy for gisting evaluation (ROUGE)-2 and an improvement of over 7% on the informativeness and conciseness metrics, as evidenced by the improved scores obtained from both automatic and human evaluations. The experimental results prove that our model can effectively select the most pertinent and meaningful content and reduce the redundancy to generate better knowledge representation.

In fuzzy clustering, the fuzzy c-means (FCM) is the most known algorithm. Several extensions and variations of FCM had been proposed in the literature. The first important extension to FCM was proposed by Gustafson and Kessel (GK). In the GK fuzzy clustering, they considered the effect of different cluster shapes except for spherical shapes by replacing the Euclidean distance of the FCM objective function with the Mahalanobis distance. The GK algorithm has become one of the most frequently used clustering algorithms. Just like FCM, the fuzziness index m is a parameter in which the value will greatly influence the performance of the GK algorithm. However, there is no theoretical work on the parameter selection for the fuzziness index m of GK. In this paper, we reveal the relation between the stable fixed points of the GK algorithm and the datasets using Jacobian matrix analysis, and then provide a theoretical base for selecting the fuzziness index m in the GK algorithm. Some experimental results verify the effectiveness of our theoretical results.

Within the bag‐of‐words (BOWs) framework, the multiple cues fusion methods provide excellent results in object and scene classification. Top‐down colour attention (CA) method is developed to use colour to guide attention by means of a top‐down category‐specific attention map. In this method, more features are taken from category‐specific colour regions where objects are more likely to be contained. In CA, the colours on the object are considered separately, so the diversity of object colours and large intra‐class colour variation make the discrimination of every colour on the object different. Object could be recognised as a collection of related colours. To enhance the object recognition capability of CA, our colour combination attention method uses mutual information and the colour combination histogram to estimate and combine the colours on the object. Results are presented on three challenging data sets, and the experiments demonstrate that the proposed feature fusion method significantly outperforms the state‐of‐the‐art methods.