Embedding Mechanism Research Articles

Construction of N-doped carbon/MnO nanoparticles/hollow carbon spheres nested structure with tunable valence state for high-performance aqueous zinc ion batteries enables 10000 cycling stability

The practical applications of manganese-based aqueous zinc ion batteries (AZIBs) have been hindered by issues such as structural collapse, dissolution of active substances, and poor conductivity. Herein, a novel N-doped carbon/MnO nanoparticles/hollow carbon spheres nested structure (HCSs@MnO@NC) is fabricated as the active cathode material for AZIBs by controlled carbothermal reduction strategy. The internal hollow structure mitigates the volume expansion of MnO and improves material stability. The Mn-O-C heterojunction structure accelerates charge transfer and enhances structural stability. By controlling the carbothermal reduction process to adjust the valence state, electrochemical determinations demonstrate that divalent manganese exhibits a higher capacitance contribution rate, which is conducive to a faster ion migration rate. In situ UV–vis is firstly employed to detect the change of Mn2+ concentration in the electrolyte, revealing a Mn2+ deposition phenomenon in addition to the ion embedding mechanism during the initial charge. Owing to these favorable structural characteristics, the prepared HCSs@MnO@NC cathode delivers a specific capacity of 320.5 mAh g−1 at 0.1 A g−1 and maintains an excellent capacity after 10,000 cycles at 3 A g−1. This work can provide a new pathway for constructing advanced AZIBs by optimizing the nanostructure of composite electrode materials.

Model can be subtle: Two important mechanisms for Social Media Popularity Prediction

Social media popularity prediction is an important channel to explore content sharing and communication on social networks. It aims to capture informative cues by analyzing multi-type data (such as user profile, image, and text) to decide the popularity of a specified post. In this paper, we divide social network users into two categories (i.e., active and inactive users) and find a dilemma in existing models: If an active user publishes the low-popularity post, the model will habitually predict the high score. On the contrary, if an inactive user provides the high-popularity post, the model still gives the low score incorrectly. Therefore, how to make the model more subtle to users is important. Comparing to existing methods that directly leverage multi-modal features for regression training, this paper stresses more on two novel mechanisms. The first method aims to prevent the over-fitting on user IDs. We propose the attribute-sensitive interactive mechanism (M1) by incorporating explicit user-attribute and post-attribute interaction. It can analyze which type of features a user cares the most and weaken the model’s dependence on user IDs. The second method aims to strengthen the influence of post content. We propose the knowledge embedding mechanism (M2) to revise the popularity scores in existing models by fusing the statistical frequency over multi-type data. Note that both mechanisms are model-agnostic, which can be applicable in any popularity prediction model. Extensive experiments conducted on the Social Media Prediction Dataset further validate the effectiveness.

A Data Augmentation Method and the Embedding Mechanism for Detection of Pulmonary Nodules on Small Samples

Abstract Lung Computed Tomography (CT) screening for pulmonary nodules provides an effective method for early diagnosis. The deep-learning-based computer-aided detection (CAD) system effectively identifies and precisely localizes suspicious pulmonary nodules in CT images, thereby significantly enhancing the accuracy and efficiency of CT diagnosis. In the medical field, the availability of medical data is limited, and research using small samples is of practical significance. By studying the data augmentation technology based on the generative model under the condition of small samples, and refining the model structure through the embedding mechanism, the accuracy and robustness of the deep learning model are explored. A 3D pixel-level statistical algorithm is proposed for the generation of pulmonary nodules. By combining simulated pulmonary nodules with healthy lung tissue, we can generate new samples of pulmonary nodules. The embedding mechanism is designed to enhance the comprehension of pixel meanings in pulmonary nodule samples by introducing latent variables. The results of the 3DVNET model with the augmentation method for pulmonary nodule detection under small sample conditions demonstrate that the proposed data augmentation method outperforms the method based on a generative adversarial network (GAN) framework, and the embedding mechanism for pulmonary nodules detection shows a significant improvement in accuracy. Conclusion: the proposed data augmentation method and embedding mechanism demonstrate significant potential in enhancing the accuracy and robustness of the model, thereby facilitating their application to various common imaging diagnostic tasks, and research using small samples is of practical significance.

Performance of Logarithmic Magnitude of Centre-Element-Based Steganography over OFDM constituted Wireless Communication Channel

Hiding crucial information is today's utmost concern and transmitting through wired or wireless medium is another attentive subject. Quality secret hiding requires robustness and better perceptual visuals of the host image. Also, a secret embedding mechanism should offer higher immunity to noises. The work proposes a robust and efficient patch-based secured steganography (P-BSS) mechanism. The stego image is transmitted over an orthogonal frequency division multiplexing (OFDM) based Wireless Communication system governed by 802.11 WLAN. The stenography framework uses center element (CE) information to embed secret bits in its 3x3 neighborhood. The strategy is based on the magnitude and natural logarithm of the center element. The statistical operation on the center element defines the position in the neighborhood from where the embedding process should be initiated. The secret binary information is then hidden using the LSB embedding approach. The stego image when transmitted over a Binary phase shift keying (BPSK) modulated OFDM wireless communication system, the stego image was immune to Additive White Gaussian Noise (AWGN) above 10 dB. The proposed logarithmic magnitude of the center-element-based steganography (LM-CEB-S) framework maintains a proper balance between imperceptibility and security.

Reversible image hiding algorithm based on adaptive embedding mechanism

A reversible blind image hiding algorithm (ReBIHA) with help of an adaptive embedding mechanism and a compressive sensing technology is presented in this paper, of which can effectively hide the secret image through a natural meaningful carrier image. In the first phase, a new key generation model (KeyGM) is built by combining random numbers and the plaintext, followed by a newly designed four dimensional chaotic map (NewCM) with stronger pseudo-randomness. Then, the generated chaotic sequence is taken to produce a measurement matrix for compressive sensing. In the second phase, the plain image is sparsely processed, confused, encrypted and compressed to get measurements. Afterward, these measurements are subjected to diffusion and binary exchange to achieve a middle cipher image (MCI). In the third stage, a meaningful carrier image is randomly selected, and integer wavelet transform (IWT) is applied to it, getting one low-frequency and three high-frequency coefficients. Then, the pixels of MCI are subjected by splitting and symmetrically shifting transformation for three element matrices. Finally, these matrices are respectively embedded into the high-frequency coefficients through an adaptive mechanism to find replaceable elements, and a final carrier image hiding secrets (CiHS) can be achieved after performing an inverse IWT. The experimental results show that the proposed algorithm can avoid the extra transmission of original carrier image. Moreover, peak signal-to-noise ratio (PSNR) of reconstructed image can surpass 32 dB at compression ratio 0.5. Especially, the proposed adaptive embedding mechanism can not only achieve blind extraction of secrets well, but also ensure the visual quality for the original carrier image.

Military Image Captioning for Low-Altitude UAV or UGV Perspectives

Low-altitude unmanned aerial vehicles (UAVs) and unmanned ground vehicles (UGVs), which boast high-resolution imaging and agile maneuvering capabilities, are widely utilized in military scenarios and generate a vast amount of image data that can be leveraged for textual intelligence generation to support military decision making. Military image captioning (MilitIC), as a visual-language learning task, provides innovative solutions for military image understanding and intelligence generation. However, the scarcity of military image datasets hinders the advancement of MilitIC methods, especially those based on deep learning. To overcome this limitation, we introduce an open-access benchmark dataset, which was termed the Military Objects in Real Combat (MOCO) dataset. It features real combat images captured from the perspective of low-altitude UAVs or UGVs, along with a comprehensive set of captions. Furthermore, we propose a novel encoder–augmentation–decoder image-captioning architecture with a map augmentation embedding (MAE) mechanism, MAE-MilitIC, which leverages both image and text modalities as a guiding prefix for caption generation and bridges the semantic gap between visual and textual data. The MAE mechanism maps both image and text embeddings onto a semantic subspace constructed by relevant military prompts, and augments the military semantics of the image embeddings with attribute-explicit text embeddings. Finally, we demonstrate through extensive experiments that MAE-MilitIC surpasses existing models in performance on two challenging datasets, which provides strong support for intelligence warfare based on military UAVs and UGVs.

Managing the transformation of foreign economic activity in an unstable macroeconomic environment

In modern conditions of chaoticization of the system of international economic relations and foreign economic activity of business entities, the tasks of improving the management of foreign economic activity of enterprises by the state and the formation of its embedded mechanisms to ensure the protection of state security in the world market are actualized. The purpose of the article is to analyze the topical issues of foreign economic activity transformation in the context of global instability. According to the author, in the system of state regulation of foreign economic activity of an enterprise, the implementation of an integrated foreign economic activity management system will be timely, which will significantly offset the problems existing in subjects of foreign economic activity related to external factors. However, it is impossible to quickly adjust to dynamically changing conditions and ensure management consistency, increase the efficiency of foreign economic activity of domestic enterprises and protect state economic security. The main goal of the transformation, according to the author, is to provide the necessary transparency and flexibility in the management system of foreign economic activity both at the state level and at the level of a specific subject of foreign economic activity, which is very important in the context of growing instability of the macroeconomic environment. Priorities in the context of growing global instability are to increase the efficiency of foreign economic activity, the competitiveness of the economy and the protection of economic security.

LncRNA FAM83H-AS1 Contributes to the Radio-resistance and Proliferation in Liver Cancer through Stability FAM83H Protein.

Liver cancer (LC) is one of China's most common malignant tumors, with a high mortality rate, ranking third leading cause of death after gastric and esophageal cancer. Recent patents propose the LncRNA FAM83H-AS1 has been verified to perform a crucial role in the progression of LC. LncRNA FAM83H-AS1 has been verified to perform a crucial role in the progression of LC. However, the concrete mechanism remains to be pending further investigation. This study aimed to explore the embedding mechanism of FAM83H-AS1 molecules in terms of radio sensitivity of LC and provide potentially effective therapeutic targets for LC therapy. Quantitative real-time PCR (qRT-PCR) was conducted to measure the transcription levels of genes. Proliferation was determined via CCK8 and colony formation assays. Western blot was carried out to detect the relative protein expression. A xenograft mouse model was constructed to investigate the effect of LncRNA FAM83H-AS1 on tumor growth and radio-sensitivity in vivo. The levels of lncRNA FAM83H-AS1 were remarkably increased in LC. Knockdown of FAM83H-AS1 inhibited LC cell proliferation and colony survival fraction. Deletion of FAM83H-AS1 increased the sensitivity of LC cells to 4 Gy of X-ray radiation. In the xenograft model, radiotherapy combined with FAM83H-AS1 silencing significantly reduced tumor volume and weight. Overexpression of FAM83H reversed the effects of FAM83H-AS1 deletion on proliferation and colony survival fraction in LC cells. Moreover, the over-expressing of FAM83H also restored the tumor volume and weight reduction caused by the knockdown of FAM83H-AS1 or radiation in the xenograft model. Knockdown of lncRNA FAM83H-AS1 inhibited LC growth and enhanced radiosensitivity in LC. It has the potential to be a promising target for LC therapy.

Tackling Few-Shot Challenges in Automatic Modulation Recognition: A Multi-Level Comparative Relation Network Combining Class Reconstruction Strategy.

Automatic Modulation Recognition (AMR) is a key technology in the field of cognitive communication, playing a core role in many applications, especially in wireless security issues. Currently, deep learning (DL)-based AMR technology has achieved many research results, greatly promoting the development of AMR technology. However, the few-shot dilemma faced by DL-based AMR methods greatly limits their application in practical scenarios. Therefore, this paper endeavored to address the challenge of AMR with limited data and proposed a novel meta-learning method, the Multi-Level Comparison Relation Network with Class Reconstruction (MCRN-CR). Firstly, the method designs a structure of a multi-level comparison relation network, which involves embedding functions to output their feature maps hierarchically, comprehensively calculating the relation scores between query samples and support samples to determine the modulation category. Secondly, the embedding function integrates a reconstruction module, leveraging an autoencoder for support sample reconstruction, wherein the encoder serves dual purposes as the embedding mechanism. The training regimen incorporates a meta-learning paradigm, harmoniously combining classification and reconstruction losses to refine the model's performance. The experimental results on the RadioML2018 dataset show that our designed method can greatly alleviate the small sample problem in AMR and is superior to existing methods.

Anomaly detection for key performance indicators by fusing self-supervised spatio-temporal graph attention networks

With the development of Artificial Intelligence for IT Operations (AIOps), numerous software and services are monitored by Key Performance Indicators (KPIs) collection components. Multivariate KPIs, as a type of time series data, are essential for effective management of the entity's service quality. In recent years, deep learning methods have made great improvements in the anomaly detection of multivariate time series; however, existing methods have not fully considered how to explicitly capture the correlation between multivariate time series in the feature dimension and temporal dimension, resulting in inevitable abnormal false positives. Therefore, this paper proposes a self-supervised multivariate KPIs anomaly detection method MAD-STA that combines graph structure learning and spatio-temporal GAT (Graph Attention Network). In the feature dimension, MAD-STA introduces a node embedding mechanism for graph structure learning and then uses the feature-oriented GAT layer to compute the graph attention coefficient to obtain the correlation between different KPIs. In the temporal dimension, MAD-STA uses the time-oriented GAT layer to compute attention weights between correlated timestamps, and the GRU-based VAE encoder captures long-term dependence to extract more comprehensive temporal feature representations. Finally, MAD-STA uses GRU-based VAE decoder to reconstruct the captured high-level features and achieves efficient anomaly detection and localization by calculating the anomaly score of multiple KPIs. Compared with the baseline methods on multiple data sets, the experimental results show that the anomaly detection accuracy of MAD-STA is better than that of the baseline method. Especially on the KPI data sets of the two server clusters of SMD and CKM, MAD-STA improves the performance and the F1 comprehensive index compared with the best baseline method. In addition, MAD-STA performs well on anomaly false positive rate and has excellent interpretability, which can be used to assist anomaly diagnosis and root cause index analysis.

SVulDetector: Vulnerability detection based on similarity using tree-based attention and weighted graph embedding mechanisms

Vulnerability detection by comparing similarities with known vulnerable code is an important method for improving code security, and is particularly effective in detecting vulnerabilities caused by code reuse. However, vulnerability detection is made difficult by the existence of some different and vulnerability-unrelated statements between codes with the same vulnerability pattern, as well as the small differences between vulnerable and fixed non-vulnerable codes. To address these challenges, we believe that more attention needs to be paid to some core syntactic and semantic information about vulnerabilities, which can help models more accurately identify vulnerable code. Hence, we propose a novel code-similarity-based vulnerability detection approach named SVulDetector. First, it contains a code representation, called Slice Composite Graphs (SCGs), which can represent rich syntactic and semantic information related to vulnerable statements while minimizing the interference from similar vulnerability irrelevant information as much as possible. Next, a tree-based attention mechanism is used to highlight certain key syntactic information in vulnerable code and fixed non-vulnerable code. Finally, SVulDetector highlights key vulnerable node information in the graph-based code representation via a weighted graph embedding mechanism. We extensively evaluated SVulDetector on an improved real-world dataset using both two-class and multi-class vulnerability detection tasks, and the proposed SVulDetector outperforms existing state-of-the-art detection methods.

Oestrogen receptor positive breast cancer and its embedded mechanism: breast cancer resistance to conventional drugs and related therapies, a review.

Traditional medication and alternative therapies have long been used to treat breast cancer. One of the main problems with current treatments is that there is an increase in drug resistance in the cancer cells owing to genetic differences such as mutational changes, epigenetic changes and miRNA (microRNA) alterations such as miR-1246, miR-298, miR-27b and miR-33a, along with epigenetic modifications, such as Histone3 acetylation and CCCTC-Binding Factor (CTCF) hypermethylation for drug resistance in breast cancer cell lines. Certain forms of conventional drug resistance have been linked to genetic changes in genes such as ABCB1, AKT, S100A8/A9, TAGLN2 and NPM. This review aims to explore the current approaches to counter breast cancer, the action mechanism, along with novel therapeutic methods endowing potential drug resistance. The investigation of novel therapeutic approaches sheds light on the phenomenon of drug resistance including genetic variations that impact distinct forms of oestrogen receptor (ER) cancer, genetic changes, epigenetics-reported resistance and their identification in patients. Long-term effective therapy for breast cancer includes selective oestrogen receptor modulators, selective oestrogen receptor degraders and genetic variations, such as mutations in nuclear genes, epigenetic modifications and miRNA alterations in target proteins. Novel research addressing combinational therapies including maytansine, photodynamic therapy, guajadiol, talazoparib, COX2 inhibitors and miRNA 1246 inhibitors have been developed to improve patient survival rates.

The biological embedding of social adversity: How adolescent housing insecurity impacts inflammation over time

Background & PurposeAdolescent housing insecurity is a dynamic form of social adversity that impacts child health outcomes worldwide. However, the means by which adolescent housing insecurity may become biologically embedded to influence health outcomes over the life course remain unclear. Therefore, we aimed to utilize life course perspectives and advanced causal inference methods to evaluate the potential for inflammation to contribute to the biological embedding of adolescent housing insecurity. Materials and methodsUsing prospective data from the Great Smoky Mountains Study, we investigated the relationship between adolescent housing insecurity and whole-blood spot samples assayed for C-reactive protein (CRP). Adolescent housing insecurity was created based on annual measures of frequent residential moves, reduced standard of living, forced separation from the home, and foster care. Annual measures of CRP ranged from 0.001 mg/L to 13.6 mg/L (median = 0.427 mg/L) and were log10 transformed to account for positively skewed values. We used g-estimation of structural nested mean models to estimate a series of conditional average causal effects of adolescent housing insecurity on CRP levels from ages 11 to 16 years and interpreted the results within life course frameworks of accumulation, recency, and sensitive periods. Principal ResultsOf the 1,334 participants, 427 [44.3 %] were female. Based on the conditional average causal effect, one exposure to adolescent housing insecurity from ages 11 to 16 years led to a 6.4 % (95 % CI = 0.69 – 12.4) increase in later CRP levels. Exposure at 14 years of age led to a 27.9 % increase in CRP levels at age 15 (95 % CI = 6.5 – 53.5). Recent exposures to adolescent housing insecurity (<3 years) suggested stronger associations with CRP levels than distant exposures (>3 years), but limited statistical power prevented causal conclusions regarding recency effects at the risk of a Type II Error. Major ConclusionsThese findings highlight inflammation—as indicated by increased CRP levels—as one potential mechanism for the biological embedding of adolescent housing insecurity. The results also suggest that adolescent housing insecurity—particularly recent, repeated, and mid-adolescent exposures—may increase the risk of poor health outcomes and should be considered a key intervention target.

Adaptive staged remaining useful life prediction of roller in a hot strip mill based on multi-scale LSTM with multi-head attention

Accurate prediction of a roller's remaining useful life (RUL) is significant for a hot strip mill to avoid major safety accidents and substantial economic losses. Since rollers’ degradation processes represent multi-stage characteristics, traditional single-stage models cannot acquire accurate RUL prediction results. Therefore, this paper proposes an adaptive staged RUL prediction method based on multi-scale long short-term memory network with multi-head self-attention (LSTM-MHA). The roller's production data and operation data are fused to construct interpretable health indicators (HI), which can represent the roller's remaining rotatable angle. Then the degradation process is divided into multiple stages and analyzed. The proposed multi-scale LSTM-MHA can adaptively update the model weights based on changes in the roller degradation stages by the multi-scale memory structure. The MHA embedding mechanism filters important temporal information for LSTM units. The trained model can predict the roller's RUL by inputting the current roller's HI. The proposed method is verified on an industrial hot strip mill roller dataset from a well-known steel company. The validation results show that the prediction accuracy of the proposed method is higher than 98.98%. Compared with existing deep learn-based methods, the proposed multi-scale LSTM-MHA method has significant advantages in roller's RUL prediction.

Fraud Feature Boosting Mechanism and Spiral Oversampling Balancing Technique for Credit Card Fraud Detection

With the flourishing of the credit card business and Internet technology, the risk of fraudulent credit card transactions is ever-increasing due to the complex information involved in the credit card business. Since the high redundancy of feature information and imbalance of class distribution in transaction data, the performance of the existing machine learning-based models for detecting credit card fraudulent transactions still needs to be improved. Therefore, it is crucial to build fraud detection models for effective feature engineering and sampling techniques. This article proposes a credit card fraud detection model incorporating a fraud feature-boosting mechanism with a spiral oversampling balancing technique (SOBT). Specifically, we present a compound grouping elimination strategy to exclude highly redundant and correlated features from the credit card transaction dataset and improve the data quality. Furthermore, we design a multifactor synchronous embedding mechanism, which combines the performance evaluation metrics of the embedding model for each feature and improves the decision-making ability of each feature for the target domain. Moreover, we propose an SOBT to balance the ratio of legitimate to fraudulent transactions, which improves the ability of the fraud detection model to distinguish legitimate from fraudulent transactions. Extensive experimental results based on two real-world datasets demonstrate that our methods can facilitate efficient credit card fraud detection and achieve better performance than state-of-the-art methods.

Fractal theory based identification model for surface crack of building structures

The existing building structures are mainly composed of concrete and masonry structures. This paper proposes a Fractal Theory-based Method (FTM) designed to efficiently detect surface cracks in both concrete and masonry building structures, addressing limitations in existing crack recognition techniques, such as weak adaptability to complex building surfaces and requiring a large amount of computational and human resource. A novel Soft Box-Counting (SBC) algorithm is proposed to calculate texture roughness information from multi-channel and multi-scale crack images, effectively mitigating the influence of non-feature information on crack identification results. The SBC algorithm is integrated into neural networks through the fractal embedding mechanism, enhancing feature analysis through visualization experiments. The resulting FTM model integrates texture roughness information from crack images and employs fractal theory to accurately classify crack regions while minimizing the similarity between crack and non-crack regions. In comparison to existing models, FTM demonstrates superior identification performance at the same computational level on benchmark datasets for surface cracks in both concrete and masonry building structures. The FTM model exhibits enhanced identification universality and higher resource utilization efficiency.

Open Knowledge Graph Link Prediction with Semantic-Aware Embedding

Link prediction in open knowledge graphs (OpenKGs) is crucial for applications like question answering and recommendation systems. Existing OpenKG models leverage the semantic information of noun phrases (NPs) to enhance the performance in the link prediction task. However, these models only extract superficial semantic information from NPs, ignoring the fact that an NP possesses diverse semantics. Furthermore, these models have not fully exploited the semantic information of the relation phrases (RPs). To address these issues, we propose a model for link prediction called Open Knowledge Graph Link Prediction with Semantic-Aware Embedding (SeAE). First, we develop an adaptive disentanglement embedding (ADE) mechanism to learn the intrinsically abundant semantics of NPs. The ADE mechanism can adaptively calculate the embedding segmentation number according to the dataset and has an ingenious method for updating embeddings. Second, we integrate the attention mechanism into the GRU encoder to obtain the distribution of importance inside RP, facilitating a more comprehensive capture of the RP’s semantic information and enhancing the model’s interpretability. Finally, we design a relation gate, which extracts the RP semantic features of tail NP from the shared edge. This gate realizes the relation constraints on entities while enhancing the interaction between entities and relations. Extensive experiments on four benchmarks demonstrate that SeAE outperforms the state-of-the-art models, resulting in improvements of approximately 5.4% and 7.4% in MRR on ReVerb45K and ReVerb45KF datasets respectively.

Neurophysiological and psychophysical references for trends in supervised VQA multimodal deep learning: An interdisciplinary meta-analysis

Leading trends in multimodal deep learning for visual-question answering include Multimodal joint-embedding model, multimodal attention-based model, and multimodal external knowledge-based model. Several mechanisms and strategies are used in these models, including representation fusion methods, co-attention mechanisms, and knowledge base retrieval mechanisms. While a variety of works have comprehensively reviewed these strategies, a key gap in research is that there is no interdisciplinary analysis that connects these mechanisms with discoveries on human. As discussions of Neuro-AI continues to thrive, it is important to consider synergies among human level investigations and ANNs, specifically for using AI to reproduce higher order cognitive functions such as multisensory integration. Thus, Present meta-analysis aimed at the reviewing and connecting neurophysiological and psychophysical references to trends in VQA multimodal deep learning, focusing on 1) Providing back-up explanations for why several strategies in VQA MMDL leads to performances that are closer to human level and 2) Using VQA MMDL as an example to demonstrate how interdisciplinary perspective may foster the development of human level AI. The result of the meta-analysis builds connections between several sub-fields: Joint embedding mechanisms and SC neurons, multimodal attention mechanism and the retro-cue effect, and external knowledge base and engram mechanisms.

Relation-oriented few-shot knowledge graph prototype networks

Knowledge graph (KG) embedding has been widely researched, but it suffers from some problems in the few-shot scenarios. The topological and semantic connection among entities cannot satisfy the assumption of samples’ independent identically distribution. Additionally, these relations between entity pairs are various and complex. But most of the previous methods are node-oriented, which is weak in modeling complex relations. In order to solve the above problems, we propose a few-shot relation prototype network (FRPN). In our method, each relation is regarded as a learning object, instead of just entity pairs’ concatenation. A relation-oriented two-channel embedding mechanism is designed to achieve multi-scale information aggregation at the entity level and the relation level respectively. For the entity level, it aggregates information at different scales according to the relation type and the neighbors under each relation. For the relation level, our model obtains each kind of relation’s prototype from the ontology and the entity layer. The multi-scale aggregation contributes to KG embedding in the few-shot scenario. Compared with existing models, our model has significantly improved the performance on both the link prediction and triple classification tasks in two few-shot datasets.

Research on gesture recognition algorithm based on MME-P3D.

A Multiscale-Motion Embedding Pseudo-3D (MME-P3D) gesture recognition algorithm has been proposed to tackle the issues of excessive parameters and high computational complexity encountered by existing gesture recognition algorithms deployed in mobile and embedded devices. The algorithm initially takes into account the characteristics of gesture motion information, integrating the channel attention (CE) mechanism into the pseudo-3D (P3D) module, thereby constructing a P3D-C feature extraction network that can efficiently extract spatio-temporal feature information while reducing the complexity of the algorithmic model. To further enhance the understanding and learning of the global gesture movement's dynamic information, a Multiscale Motion Embedding (MME) mechanism is subsequently designed. The experimental findings reveal that the MME-P3D model achieves recognition accuracies reaching up to 91.12% and 83.06% on the self-constructed conference gesture dataset and the publicly available Chalearn 2013 dataset, respectively. In comparison with the conventional 3D convolutional neural network, the MME-P3D model demonstrates a significant advantage in terms of parameter count and computational requirements, which are reduced by as much as 82% and 83%, respectively. This effectively addresses the limitations of the original algorithms, making them more suitable for deployment on embedded and mobile devices and providing a more effective means for the practical application of hand gesture recognition technology.