Relational Features Research Articles

Gaia Data Release 3

Context. The third Gaia data release provides photometric time series covering 34 months for about 10 million stars. For many of those stars, a characterisation in Fourier space and their variability classification are also provided. This paper focuses on intermediate- to high-mass (IHM) main sequence pulsators (M ≥ 1.3 M⊙) of spectral types O, B, A, or F, known as β Cep, slowly pulsating B (SPB), δ Sct, and γ Dor stars. These stars are often multi-periodic and display low amplitudes, making them challenging targets to analyse with sparse time series. Aims. We investigate the extent to which the sparse Gaia DR3 data can be used to detect OBAF-type pulsators and discriminate them from other types of variables. We aim to probe the empirical instability strips and compare them with theoretical predictions. The most populated variability class is that of the δ Sct variables. For these stars, we aim to confirm their empirical period-luminosity (PL) relation, and verify the relation between their oscillation amplitude and rotation. Methods. All datasets used in this analysis are part of the Gaia DR3 data release. The photometric time series were used to perform a Fourier analysis, while the global astrophysical parameters necessary for the empirical instability strips were taken from the Gaia DR3 gspphot tables, and the v sin i data were taken from the Gaia DR3 esphs tables. The δ Sct PL relation was derived using the same photometric parallax method as the one recently used to establish the PL relation for classical Cepheids using Gaia data. Results. We show that for nearby OBAF-type pulsators, the Gaia DR3 data are precise and accurate enough to pinpoint them in the Hertzsprung-Russell (HR) diagram. We find empirical instability strips covering broader regions than theoretically predicted. In particular, our study reveals the presence of fast rotating gravity-mode pulsators outside the strips, as well as the co-existence of rotationally modulated variables inside the strips as reported before in the literature. We derive an extensive period–luminosity relation for δ Sct stars and provide evidence that the relation features different regimes depending on the oscillation period. We demonstrate how stellar rotation attenuates the amplitude of the dominant oscillation mode of δ Sct stars. Conclusions. The Gaia DR3 time-series photometry already allows for the detection of the dominant (non-)radial oscillation mode in about 100 000 intermediate- and high-mass dwarfs across the entire sky. This detection capability will increase as the time series becomes longer, allowing the additional delivery of frequencies and amplitudes of secondary pulsation modes.

Learning Social Spatio-Temporal Relation Graph in the Wild and a Video Benchmark.

Social relations are ubiquitous and form the basis of social structure in our daily life. However, existing studies mainly focus on recognizing social relations from still images and movie clips, which are different from real-world scenarios. For example, movie-based datasets define the task as the video classification, only recognizing one relation in the scene. In this article, we aim to study the problem of social relation recognition in an open environment. To close the gap, we provide the first video dataset collected from real-life scenarios, named social relation in the wild (SRIW), where the number of people can be huge and vary, and each pair of relations needs to be recognized. To overcome new challenges, we propose a spatio-temporal relation graph convolutional network (STRGCN) architecture, utilizing correlative visual features to recognize social relations intuitively. Our method decouples the task into two classification tasks: person-level and pair-level relation recognition. Specifically, we propose a person behavior and character module to encode moving and static features in two explicit ways. Then we take them as node features to build a relation graph with meaningful edges in a scene. Based on the relation graph, we introduce the graph convolutional network (GCN) and local GCN to encode social relation features which are used for both recognitions. Experimental results demonstrate the effectiveness of the proposed framework, achieving 83.1% and 40.8% mAP in person-level and pair-level classification. Moreover, the study also contributes to the practicality in this field.

T-GAE: A Timespan-aware Graph Attention-based Embedding Model for Temporal Knowledge Graph Completion

Temporal knowledge graphs (TKGs) often suffer from incompleteness, leading to an important research issue: Temporal Knowledge Graph Completion (TKGC). Knowledge Graph Embedding (KGE) methods have proven to be effective in solving this issue. However, most of them handle triples independently and do not capture complex information embedded in the neighborhood topology of central entities. To this end, we propose a Timespan-awareGraphAttention-basedEmbedding Model named T-GAE to tackle the TKGC task. To the best of our knowledge, T-GAE is the first KGE model in which Graph-Attention-Networks (GATs) and Long Short-Term Memory (LSTM) Networks are simultaneously applied to the TKGC task. In essence, our model is an Encoder-Decoder architecture, where the encoder consists of an LSTM network and a GAT network. Firstly, we employ LSTM layers to learn new time-aware relational embeddings to incorporate time information. Then, we utilize these time-aware relational embedding and GATs considered as neighborhood aggregators to learn the entity and relational features of the central entity neighborhoods. Thus, T-GAE can capture the interaction features between multi-relational facts and the abundant temporal information in TKGs. As for the decoder, we choose the ConvKB model, which is essentially a scoring function. Our experiments demonstrate that T-GAE significantly outperforms most of the existing baseline methods for TKGC in terms of MRR and Hit@1/3/10.

Instance Segmentation of Irregular Deformable Objects for Power Operation Monitoring Based on Multi-Instance Relation Weighting Module

Electric power operation is necessary for the development of power grid companies, where the safety monitoring of electric power operation is difficult. Irregular deformable objects commonly used in electrical construction, such as safety belts and seines, have a dynamic geometric appearance which leads to the poor performance of traditional detection methods. This paper proposes an end-to-end instance segmentation method using the multi-instance relation weighting module for irregular deformable objects. To solve the problem of introducing redundant background information when using the horizontal rectangular box detector, the Mask Scoring R-CNN is used to perform pixel-level instance segmentation so that the bounding box can accurately surround the irregular objects. Considering that deformable objects in power operation workplaces often appear with construction personnel and the objects have an apparent correlation, a multi-instance relation weighting module is proposed to fuse the appearance features and geometric features of objects so that the relation features between objects are learned end-to-end to improve the segmentation effect of irregular objects. The segmentation mAP on the self-built dataset of irregular deformable objects for electric power operation workplaces reached up to 44.8%. With the same 100,000 training rounds, the bounding box mAP and segmentation mAP improved by 1.2% and 0.2%, respectively, compared with the MS R-CNN. Finally, in order to further verify the generalization performance and practicability of the proposed method, an intelligent monitoring system for the power operation scenes is designed to realize the actual deployment and application of the proposed method. Various tests show that the proposed method can segment irregular deformable objects well.

"All my co-workers are good people, but…": Collaboration dynamics between frontline workers in rural Uttar Pradesh, India.

Multi-sectoral collaboration has been identified as a critical component in a wide variety of health and development initiatives. For India's Integrated Child Development Services (ICDS) scheme, which serves over 100 million people annually across more than one million villages, a key point of multi-sectoral collaboration - or "convergence", as it is often called in India - is between the three frontline worker cadres jointly responsible for delivering essential maternal and child health and nutritional services throughout the country: the Accredited Social Health Activist (ASHA), Anganwadi worker (AWW), and auxiliary nurse midwife (ANM), or "AAA" workers. Despite the long-recognized importance of collaboration within this triad, there has been relatively little documentation of what this looks like in practice and what is needed in order to improve it. Informed by a conceptual framework of collaborative governance, this study applies inductive thematic analysis of in-depth interviews with 18 AAA workers and 6 medical officers from six villages across three administrative blocks in Hardoi district of Uttar Pradesh state to identifiy key elements of collaboration. These are grouped into three broad categories: organizational (including interdependence, role clarity, guidance/support, and resource availability); relational (interpersonal, conflict resolution); and personal (flexibility, diligence, locus of control). These findings underscore the importance of personal and relational collaboration features, which are underemphasized in India's ICDS, the largest of its kind globally, and in the multi-sectoral collaboration literature more broadly - both of which place greater emphasis on organizational aspects of collaboration. These findings are largely consistent with prior studies, but are notably different in that they highlight the importance of flexibility, locus of control, and conflict resolution in collaborative relationships, all of which relate to one's ability to adapt to unexpected obstacles and find mutually workable solutions with colleagues. From a policy perspective, supporting these key element of collaboration may involve giving frontline workers more autonomy in how they get the work done, which may in some cases be impeded by additional training to reinforce worker role delineation, closer monitoring, or other top-down efforts to push greater convergence. Given the essential role that frontline workers play in multi-sectoral initiatives in India and around the world, there is a clear need for policymakers and managers to understand the elements affecting collaboration between these workers when designing and implementing programs.

A Parallel Fusion Graph Convolutional Network for Aspect-Level Sentiment Analysis

Sentiment analysis has always been an important basic task in the NLP field. Recently, graph convolutional networks (GCNs) have been widely used in aspect-level sentiment analysis. Because GCNs have good aggregation effects, every node can contain neighboring node information. However, in previous studies, most models used only a single GCN to learn contextual information. The GCN relies on the construction method of the graph, and a single GCN will cause the model to focus on a certain relationship of nodes that depends on the construction method and ignore other information. In addition, when the GCN aggregates node information, it cannot determine whether the aggregated information is useful, so it will inevitably introduce noise. We propose a model that fuses two parallel GCNs to learn different relational features between sentences at the same time, and we add a gate mechanism to the GCN to filter the noise introduced by the GCN when aggregating information. Finally, we validate our model on public datasets, and the experiments show that compared to state-of-the-art models, our model performs better.

Hypercomplex context guided interaction modeling for scene graph generation

Intuitively, humans can consciously and subjectively attend to the interactions between objects, and thus infer reasonable visual relations. However, mainstream approaches of Scene Graph Generation (SGG) strive to alleviate the long-tailed distribution problem with various complicated re-weighting strategies, where a simple concatenation of the refined object features is treated as the final representation of visual relations. In spite of their remarkable progress, such an operation overlooks the importance of interaction on relation recognition. To tackle the problem, this work devises a hyperComplex- Context guided Interaction Modeling (CCIM for short) plug-in, which can be successfully assimilated by the existing methods for performance improvement. Specifically, we first extract the contextual relation feature determined by the constraint relation≈union(head,tail)−headobject−tailobject. Then, we encode the features of relations and objects into hypercomplex space, with three imaginary components, to learn more expressive representations for SGG. Next, guided by the context, we can capture the interaction between a head or tail object and their relation through the Hamilton product. We further reinforce the interaction between enhanced hypercomplex-valued representations of the two entities with Quaternion inner product. At last, the concatenation of all components from the learned hypercomplex feature is adopted as our final relation representation. Extensive experiments on the popular benchmark Visual Genome in various existing approaches demonstrate the effectiveness and generalization of our proposed model-agnostic method under comprehensive evaluation metrics.

A coarse aggregate gradation detection method based on 3D point cloud

Coarse aggregate plays the role of skeleton and accounts for the highest proportion in asphalt pavement, and its particle size and gradation distribution have an important impact on pavement performance. Traditional gradation detection methods by square sieves are inefficient and labor-intensive. This paper is a study of automated gradation detection methods based on 3D point clouds and deep learning. A line structured light based acquisition system was built to collect aggregate point cloud data, and a multi-threshold segmentation method is proposed to segment each aggregate from the background. To solve the problem of high aggregate complexity and low discrimination between different particle sizes, we propose an aggregate particle size classification model ISRNet (Improved Structure Relation Network) with pointnet. Based on SA (Set Abstraction), we add adaptive sampling (AS) after Farthest Point Sampling (FPS) to shift the sampling points adaptively. In addition, for the characteristic that the aggregate particle size is affected by different local structures, we propose to add ISRN after SA to learn the relation features. For the quality prediction problem, a GA-LightGBM model was developed with nine 3D features as input and parameter optimization inspired by genetic algorithms. In addition, the effects of different acquisition parameters on the model are discussed. Compared with other classification algorithms, the accuracy of the proposed classification method achieved 99.12% in the range of 0–10 mm and 98.89% in the range of 10–20 mm. Moreover, the maximum error of aggregate gradation evaluation is 1.63%, which can meet the industrial requirement. It proves the advantages of 3D point cloud in determining the aggregate particle size and gradation, improving the automatic detection efficiency and saving labor costs.

Knowledge graph representation learning with simplifying hierarchical feature propagation

Graph neural networks (GNN) have emerged as a new state-of-the-art for learning knowledge graph representations. Although they have shown impressive performance in recent studies, how to efficiently and effectively aggregate neighboring features is not well designed. To tackle this challenge, we propose the simplifying heterogeneous graph neural network (SHGNet), a generic framework that discards the two standard operations in GNN, including the transformation matrix and nonlinear activation. SHGNet, in particular, adopts only the essential component of neighborhood aggregation in GNN and incorporates relation features into feature propagation. Furthermore, to capture complex structures, SHGNet utilizes a hierarchical aggregation architecture, including node aggregation and relation weighting. Thus, the proposed model can treat each relation differently and selectively aggregate informative features. SHGNet has been evaluated for link prediction tasks on three real-world benchmark datasets. The experimental results show that SHGNet significantly promotes efficiency while maintaining superior performance, outperforming all the existing models in 3 out of 4 metrics on NELL-995 and in 4 out of 4 metrics on FB15k-237 dataset.

A joint training network for learning more distinguishable relation features in relation classification

Relation classification is an important task in natural language processing, which aims to predict the semantic relation between a given entity pair in a sentence. There are datasets, like TACRED, that contain a large number of “no_relation” type samples. Most existing methods treat “no_relation” and normal relation types equally, and directly apply the softmax function over all relation types. In this paper, we propose a novel joint training network to learn more distinguishable relation features for relation classification. Specially, we convert the original multi-class classification problem into two joint optimized modules, binary classification of whether a relation is “no_relation” and multi-class classification of normal relation types. To further differentiate between similar normal relation types, we introduce a self-supervised contrastive learning method to learn more distinguishable features for them. We jointly optimize the above modules. Experimental results agree well with our design intention and demonstrate that our joint training network not only achieves superior performance against existing competitive models, but also is robust to “no_relation” problem.

The Old Bulgarian Demonstrative Pronoun сь as a Subordinate Element in the Verb Group and its Equivalents in Old English

This paper presents a model for the analysis of pronouns based on two types of characterizations – semantic-grammatical and communicative-pragmatic ones. Each of them consists of a set of categorial and relational features. The semantic-grammatical and communicative-pragmatic characterizations determine the types of deixis expressed by pronouns. The model is tested on the demonstrative pronoun сь used as a subordinate element in the verb group. Excerpted material from Old Bulgari-an Gospel texts is studied in comparison with Old English, Greek and Latin translations.

Process Monitoring Based on Temporal Feature Agglomeration and Enhancement

Dear Editor, This letter proposes a process-monitoring method based on temporal feature agglomeration and enhancement, in which a novel feature extractor called contrastive feature extractor (CFE) extracts the temporal and relational features among process parameters. Then the feature representations are enhanced by maximizing the separation among different classes while minimizing the scatter within each class.

Detection of Generative Linguistic Steganography Based on Explicit and Latent Text Word Relation Mining Using Deep Learning

Covert communication channels can be easily constructed using text steganography based on social media. Offenders can easily utilize these channels to engage in various criminal activities, which brings great challenges in maintaining the security of cyberspace. Among the text information hiding methods, generative linguistic steganography poses the biggest threat to network security because it does not need the original carrier and has high embedding efficiency. The existing generative linguistic steganalysis methods fail to deeply mine text word relation, hence the detection performance is relatively unsatisfactory. In this paper, we prove that there is explicit and latent steganography-sensitive text word relation. Based on this, we propose a generative linguistic steganalysis method based on Explicit and Latent text word relation Mining, named ELM. Firstly, we employ a distributed readin module to convert words into real number vectors. Then, MRA (Mining Relation by Attentions) is proposed to mine the explicit and latent text word relation. Finally, global adaptive classification module is presented to exploit the mined relation feature to predict whether secret information is embedded in the current text segment. Experimental results demonstrate that the detection performance of ELM is better than the existing generative linguistic steganalysis methods.

Eater-oriented knowledge framework for reducing salt and dietary sodium intake (scoping review).

Salt and dietary sodium are ubiquitously present in daily food practices and, at the same time, reducing salt intake presents an important public health issue. Given such an ambivalent position of salt in human diet, we argue that public health guidelines toward dietary sodium reduction require an eater-oriented knowledge framework. In this article we are making the first steps toward a flexible interdisciplinary database which would include nutritional, socio-economic, cultural, material, and socio-psychological determinants of salt consumption for comprehensive public health campaigns. We employ an explorative scoping review of academic articles and reports, limiting our review to the original data on salt or sodium consumption published in English or French between 2000 and 2022. We describe salt consumption as research object, identifying its representation in different research fields, data sources, methodologies, samples, and links with nutritional recommendations. We synthesize existing approaches via four eater-oriented categories: Socio-demographic and cultural descriptors of salt consumers; Knowledge, attitudes, and beliefs on nutritional norms; Salt practices associated with dietary or medical regimes; Salt materialities: interactions and contexts. In each category, we identify the dominant relational features, i.e., what kind of 'eater-salt' relation is being put forward. We thus build an interdisciplinary documentary base of dietary sodium consumption factors. We discuss the results, suggesting that comprehensive nutritional policies for global salt reduction require interdisciplinary eater-oriented data frameworks.

An end-to-end neural framework using coarse-to-fine-grained attention for overlapping relational triple extraction

Abstract In recent years, the extraction of overlapping relations has received great attention in the field of natural language processing (NLP). However, most existing approaches treat relational triples in sentences as isolated, without considering the rich semantic correlations implied in the relational hierarchy. Extracting these overlapping relational triples is challenging, given the overlapping types are various and relatively complex. In addition, these approaches do not highlight the semantic information in the sentence from coarse-grained to fine-grained. In this paper, we propose an end-to-end neural framework based on a decomposition model that incorporates multi-granularity relational features for the extraction of overlapping triples. Our approach employs an attention mechanism that combines relational hierarchy information with multiple granularities and pretrained textual representations, where the relational hierarchies are constructed manually or obtained by unsupervised clustering. We found that the different hierarchy construction strategies have little effect on the final extraction results. Experimental results on two public datasets, NYT and WebNLG, show that our mode substantially outperforms the baseline system in extracting overlapping relational triples, especially for long-tailed relations.

Knowledge Extraction From National Standards for Natural Resources

National standards for natural resources (NSNR) plays an important role in promoting efficient use of China's natural resources, which sets standards for many domains such as marine and land resources. Its revision is difficult since standards in different domains may overlap or conflict. To facilitate the revision of NSNR, this paper extracts structural knowledge from the NSNR files to assist its revision. NSNR files are in multi-domain texts, where the traditional knowledge extraction methods could fall short in recalling multi-domain entities. To address this issue, this paper proposes a knowledge extraction method for multi-domain texts, including sub-domain relation discovery (SRD) and domain semantic features fusion (DSFF) module. SRD splits NSNR into sub-domains to facilitate the relation discovery. DSFF integrates relation features in the conditional random field (CRF) model to improve the capability of multi-domain entity recognition. Experimental results demonstrate that the proposed method could effectively extract structural knowledge from NSNR.

HyGGE: Hyperbolic graph attention network for reasoning over knowledge graphs

Recently, hyperbolic embedding has successfully demonstrated its superiority over Euclidean analogues in representing hierarchical data. As the scale-free network that usually exhibits rich hierarchical structures, knowledge graphs naturally become a field where hyperbolic embedding shows its talents. Furthermore, hyperbolic embedding is also an expected solution to comprehensively reproduce the semantic features and underlying structures of KGs in the embedding space, which will significantly optimize the interpretability and performance of embedding models. However, most of the several existing hyperbolic studies only individually learn semantic information indicated by triples individually, making the embedding space relatively one-sided and simplified. In addition, many issues that limit reasoning performance are still ignored and unresolved in the context of hyperbolic geometry, like the response to complex relations and relation patterns. Motivated by these concerns, we propose the hyperbolic embedding model for KG reasoning, HyGGE. It is based on an innovative hyperbolic graph attention network. Furthermore, the response to complex relations, which is a well-known problem that constrains reasoning performance is also discussed in HyGGE. On the one hand, the focus on neighborhood structures and relation features makes up for the singularity that the embedding space is completely induced by triples individually, thereby optimizing the expressiveness of the embedding space. On the other hand, they cooperate with the effect of hyperbolic geometry to capture hierarchical features contained in local structures, and thus giving the hyperbolic embedding a fuller play to its advantages. Extensive experiments have validated the effectiveness and advantages of HyGGE.

SSD-KD: A self-supervised diverse knowledge distillation method for lightweight skin lesion classification using dermoscopic images.

Skin cancer is one of the most common types of malignancy, affecting a large population and causing a heavy economic burden worldwide. Over the last few years, computer-aided diagnosis has been rapidly developed and make great progress in healthcare and medical practices due to the advances in artificial intelligence, particularly with the adoption of convolutional neural networks. However, most studies in skin cancer detection keep pursuing high prediction accuracies without considering the limitation of computing resources on portable devices. In this case, the knowledge distillation (KD) method has been proven as an efficient tool to help improve the adaptability of lightweight models under limited resources, meanwhile keeping a high-level representation capability. To bridge the gap, this study specifically proposes a novel method, termed SSD-KD, that unifies diverse knowledge into a generic KD framework for skin disease classification. Our method models an intra-instance relational feature representation and integrates it with existing KD research. A dual relational knowledge distillation architecture is self-supervised trained while the weighted softened outputs are also exploited to enable the student model to capture richer knowledge from the teacher model. To demonstrate the effectiveness of our method, we conduct experiments on ISIC 2019, a large-scale open-accessed benchmark of skin diseases dermoscopic images. Experiments show that our distilled MobileNetV2 can achieve an accuracy as high as 85% for the classification tasks of 8 different skin diseases with minimal parameters and computing requirements. Ablation studies confirm the effectiveness of our intra- and inter-instance relational knowledge integration strategy. Compared with state-of-the-art knowledge distillation techniques, the proposed method demonstrates improved performance. To the best of our knowledge, this is the first deep knowledge distillation application for multi-disease classification on the large-scale dermoscopy database. Our codes and models are available at https://github.com/enkiwang/Portable-Skin-Lesion-Diagnosis.

목표 기반 시각적 이동 작업을 위한 시각적 맥락정보 임베딩과 교착상태 처리

In this paper, we propose a novel deep neural network-based agent model, (VCENet), for performing target-driven visual navigation tasks. Most recent agent models for visual navigation try to recognize the real-time task context by using only objects and their relationships detected from RGB input images. However, such object-oriented visual context does not contain detailed information about the background scene, which is important for ground navigation. Moreover, it may result in a misrecognized context due to because of errors in of the object and relation detectors. To overcome these problems, the proposed VCENet model represents the real-time task context by using as the appearance and geometric features of the background scene extracted from RGB-D input images well as the object relation features derived through from graph embedding. Many existing models, which learn an action policiesy based upon on reinforcement, learning, do not provide any reward functions to avoid deadlocks ahead during of navigation or any effective deadlock recovery policyy learning mechanisms to escape the deadlocks. To address these problems, the proposed VCENet model provides not only reward functions for deadlock avoidance and recovery, but also a deadlock recovery policy based on imitation learning. Conducting a variety of various experiments in photo-realistic virtual indoor environments provided by the 3D simulator AI2THOR, we made sure of demonstrated the superiority of the proposed VCENet model.

Deep Featured Adaptive Dense Net Convolutional Neural Network Based Cardiac Risk Prediction in Big Data Healthcare Environment

In recent days, cardiac vascular disease has been one of the deadliest health-affecting factors causing sudden death. So, the importance of early risk prediction through feature analysis has become a big problem in data analysis because more nonlinear time series data increase the feature dimension. Irrelevant feature dimension scaling affects the prediction accuracy and leads to classification inaccuracy. To resolve this problem, we propose an Enhanced Healthcare data analysis model for cardiac data prediction using an adaptive Deep Featured Adaptive Convolution Neural Network for early risk identification. Initially, the preprocessing was augmented to formalize the time series data collected from the CVD-DS dataset. Then the feature evaluation was carried out with the Relative Subset Clustering (RSC) approach. The Cardiac Deficiency Prediction rate (CDPr) was estimated to identify the relational feature to subset margins. Based on the CDPr weight the feature is extracted using Cross-Over Mutual Scaling Feature Selection Model (CMSFS). The selected features get with a deep neural classifier based on logical neurons. They are then constructed into a Dense Net Convolution Neural Network (DN-CNN) classifier to feed forward the feature values and predict the Disease Affection Rate (DAR) by class category. The proposed system produces high prediction accuracy in classification, precision, and recall rate to support premature treatment for early cardiac disease risk prediction..