Adjacency Relations Research Articles

A virtual sensor network for pressure distribution inside a multi-zone building based on spatial adjacency relationships and multivariate adaptive regression spline

Airflows within multi-zone buildings cause variations in environmental parameters both vertically and horizontally. Although it is difficult to monitor the airflows through existing sensors, pressure monitoring allows the identification of airflows, since airflows across zones are induced by pressure differences. This study proposes a data-driven method for establishing a virtual pressure sensor network within a multi-zone building. Three scenarios with various sample sizes are applied to evaluate their impacts on the analyzed results, and scenario 3 is identified as the most suitable database for creating a virtual sensor network. The single-sided or multi-sided spatial adjacency relationships are determined for each zone, and the MARS model is then applied to develop the relationships among the variables, with key physical sensors designated by the frequency analysis of independent variables. The final sensor network comprises four key physical sensors, thirteen direct virtual sensors, and five indirect virtual sensors. All virtual sensors can achieve excellent predicted results with high R2 values and low MAE and RMSE, with mean values of 0.997 Pa and 1.269 Pa, respectively. Notably, the difference in accuracy between direct and indirect virtual sensors is influenced by spatial adjacency relationships. This study demonstrates a reliable virtual sensor network for predicting pressure variations and provides a reference for the identification of airflows and the efficient adjustment of HVAC systems.

Dynamic directed graph convolution network based ultra‐short‐term forecasting method of distributed photovoltaic power to enhance the resilience and flexibility of distribution network

AbstractAccurately forecasting regional distributed photovoltaic (DPV) power is crucial in mitigating the negative impact of high DPV penetration on the reliability and resilience of the distribution network. However, most of the current photovoltaic power forecasting methods suffer from two key problems: (1) ignoring the asymmetric influence relationship among DPV sites; (2) lack of consideration of dynamic spatiotemporal correlation among DPV sites. As a result, these methods are unable to fully adapt to the characteristics of DPV, making it challenging to directly apply the existing forecasting methods to improve the accuracy of DPV power forecasting. To conquer this limitation, a dynamic directed Graph Convolution Neural Network (DDGCN) is applied to regional DPV ultra‐short‐term power forecasting. Unlike the conventional Graph Convolution Neural Network (GCN) based forecasting methods, the proposed method improves GCN to process the directed graph. On this basis, to capture the dynamic and directed adjacency relationship among graph nodes, a temporal attention mechanism is proposed and combined with the directed GCN model. In this way, the dynamic and asymmetric/directed relationships among DPV sites can be taken into account. It is worth noting that the DPVs’ adjacency relationship can be constructed without any prior knowledge by end‐to‐end model training. The simulation experiment proves that the prediction accuracy can be further improved by taking into account the dynamic directed relationship among the sites via a real DPV power dataset.

Spatio-temporal modeling for estimation of bigeye tuna catch in the presence of pandemic-related data loss using parametric adjacency structures

Conditionally autoregressive (CAR) space-time models have wide applicability because spatial association structure can be captured through an adjacency matrix, without expressly relying on distance information. However, in the traditional spatio-temporal CAR (STCAR), the adjacency matrix is based on adjacency relationships, which can be problematic in practical applications where those relationships are dependent on non-spatial considerations. A new formulation of a parametric structure for the adjacency structure was developed, offering a functional and flexible solution to this problem. This method is applied to estimate bigeye tuna catch for the purse-seine fishery in the eastern Pacific Ocean for 2020–2021. During the COVID-19 pandemic, collection of some data types for this fishery was severely negatively impacted, resulting in non-random loss of data. To mitigate this, a STCAR model was developed, combining multiple sources of spatio-temporal data to obtain enhanced prediction results. The objective was to produce statistically similar catch estimates to the historical time series on which current fisheries management is based, thereby minimizing bias due to a change in estimation methodology. The traditional CAR formulation assumes uniform weight in defining spatial dependence between neighboring areas when they share boundaries or corners, and no weight when the areal units do not share boundaries or corners. The new parametric approach uses expert opinions to define weighted adjacency in the formulation where the ‘weight’ is based on how substitutions rules apply when data is missing and areas can be ‘adjacent’ in terms of characteristics rather than simply spatially. The new parametric formulation of the adjacency matrix performed best, as compared to traditional formulations, was asymptotic as the parameter got smaller. This parametric specification of the adjacency matrix, along with an AR (1) structure in the STCAR, performed robustly when a sensitivity analysis was carried out by simulating non-random pandemic-like data-loss for pre-pandemic years.

Query Path Generation via Bidirectional Reasoning for Multihop Question Answering From Knowledge Bases

Multi-hop question answering from knowledge bases (KBQA) is a hot research topic in natural language processing. Recently, graph neural network-based (GNN-based) methods have achieved promising results as the KB can be organized as a knowledge graph (KG). However, they often suffered from the sparsity of the KG which was detrimental to the structure-encoding and reasoning capabilities of GNN. Specifically, a KG is a sparse graph linked by directed relations and previous studies have paid scant attention to the directional characteristic of relations in the KG, limiting the patterns of relation path that GNN-based approaches could resolve. This study proposes a bidirectional recurrent graph neural network (BRGNN) to tackle these difficulties. To model the bidirectional information of relations, all adjacent relations of an entity are grouped by their directions, and they are separately aggregated into the entity representation in outward and inward directions. For the reasoning process, BRGNN simultaneously considers the neighbor relations in both directions to cover more patterns of relation paths and improve the recall of answers. Extensive experiments on three benchmarks: WebQuestionsSP, ComplexWebQuestions and MetaQA, verify that BRGNN can answer more questions by taking into account the directional information, and it is competitive to all state-of-the-art approaches.

Multi-robot raster map fusion without initial relative position

PurposeThe purpose of this study is to solve the problem of an unknown initial position in a multi-robot raster map fusion. The method includes two-dimensional (2D) raster maps and three-dimensional (3D) point cloud maps.Design/methodology/approachA fusion method using multiple algorithms was proposed. For 2D raster maps, this method uses accelerated robust feature detection to extract feature points of multi-raster maps, and then feature points are matched using a two-step algorithm of minimum Euclidean distance and adjacent feature relation. Finally, the random sample consensus algorithm was used for redundant feature fusion. On the basis of 2D raster map fusion, the method of coordinate alignment is used for 3D point cloud map fusion.FindingsTo verify the effectiveness of the algorithm, the segmentation mapping method (2D raster map) and the actual robot mapping method (2D raster map and 3D point cloud map) were used for experimental verification. The experiments demonstrated the stability and reliability of the proposed algorithm.Originality/valueThis algorithm uses a new visual method with coordinate alignment to process the raster map, which can effectively solve the problem of the demand for the initial relative position of robots in traditional methods and be more adaptable to the fusion of 3D maps. In addition, the original data of the map can come from different types of robots, which greatly improves the universality of the algorithm.

Logic-based benders decomposition for wildfire suppression

In this paper we consider a network interdiction problem where the follower is characterised by a spreading phenomenon. As a case study we consider in detail a wildfire suppression problem. The landscape is modelled as a directed graph, with nodes representing regions of the landscape, and arcs representing adjacency relationships. The fire spread is modelled using the minimum travel time principle. The goal is to locate fire suppression resources in a landscape in order to minimise the total area burned. As previous exact algorithms in this space are intractable on reasonable instances, the literature focuses on heuristics. In this paper we propose a new exact algorithm utilising logic-based Benders decomposition. We benchmark the new approach against mixed-integer programming and metaheuristic approaches. Our approach is able to quickly solve instances from the literature. We are also able to solve larger instances to optimality in a reasonable amount of time.

Guidewire Endpoint Detection Based on Pixel Adjacent Relation in Robot-assisted Cardiovascular Interventions.

Visualization of endovascular tools like guidewire and catheter is essential for procedural success of endovascular interventions. This requires tracking the tool pixels and motion during catheterization; however, detecting the endpoints of the endovascular tools is challenging due to their small size, thin appearance, and flexibility. As this still limit the performances of existing methods used for endovascular tool segmentation, predicting correct object location could provide ways forward. In this paper, we proposed a neighborhood-based method for detecting guidewire endpoints in X-ray angiograms. Typically, it consists of pixel-level segmentation and a post-segmentation step that is based on adjacency relationships of pixels in a given neighborhood. The latter includes skeletonization to predict endpoint pixels of guidewire. The method is evaluated with proprietary guidewire dataset obtained during in-vivo study in six rabbits, and it shows a high segmentation performance characterized with precision of 87.87% and recall of 90.53%, and low detection error with a mean pixel error of 2.26±0.14 pixels. We compared our method with four state-of-the-art detection methods and found it to exhibit the best detection performance. This neighborhood-based detection method can be generalized for other surgical tool detection and in related computer vision tasks.Clinical Relevance- The proposed method can be provided with better tool tracking and visualization systems during robot-assisted intravascular interventional surgery.

Multi-view expressive graph neural networks for 3D CAD model classification

The creation of effective content-based retrieval systems for the 3D models of engineering components created by Mechanical Computer Aided Design (MCAD) systems has been a subject of academic investigation since the 1990 s. Recently some of the most promising results have been reported by researchers using Deep Neural Nets (DNN) to classify industrial parts represented by collections of images generated from different viewpoints. Known as Multi-View Convolutional Neural Network (MVCNN) these systems have extended the architectures developed for 2D images analysis to handle 3D data by using a series of pictures rendered from different viewpoints. In 2016 Monti et al. used Graph Neural Networks (GNN) to classify the superpixel images for the first time and reported results which suggested that the approach could produce good classification accuracy (an observation confirmed by the application of GNNs to 2D image datasets of common objects such as MNIST and Cifar). To investigate the potential for GNNs to reason about spatial adjacency relationships this paper reports, for the first time, the application of GNNs to classifications of 3D MCAD models of mechanical components. Viewing GNN as a convolution operator, parallel GNN can be applied to multiple views of a 3D shape in the same way as CNN is structured in a MVCNN. After outlining the architecture implemented to support MVGNNs the impact of various hyperparameters on the expressive power are explored. When optimised these hyperparameters (located in both intra-layer and inter-layer structures) combined with the multi-view architecture produced a classification accuracy superior to previous methods. The results also suggest that GNNs have the potential to produce fast, accurate classification systems for 3D MCAD data using significantly smaller datasets than those used for transfer learning in MVCNNs.

RWEN-TTS: Relation-Aware Word Encoding Network for Natural Text-to-Speech Synthesis

With the advent of deep learning, a huge number of text-to-speech (TTS) models which produce human-like speech have emerged. Recently, by introducing syntactic and semantic information w.r.t the input text, various approaches have been proposed to enrich the naturalness and expressiveness of TTS models. Although these strategies showed impressive results, they still have some limitations in utilizing language information. First, most approaches only use graph networks to utilize syntactic and semantic information without considering linguistic features. Second, most previous works do not explicitly consider adjacent words when encoding syntactic and semantic information, even though it is obvious that adjacent words are usually meaningful when encoding the current word. To address these issues, we propose Relation-aware Word Encoding Network (RWEN), which effectively allows syntactic and semantic information based on two modules (i.e., Semantic-level Relation Encoding and Adjacent Word Relation Encoding). Experimental results show substantial improvements compared to previous works.

LORE: Logical Location Regression Network for Table Structure Recognition

Table structure recognition (TSR) aims at extracting tables in images into machine-understandable formats. Recent methods solve this problem by predicting the adjacency relations of detected cell boxes, or learning to generate the corresponding markup sequences from the table images. However, they either count on additional heuristic rules to recover the table structures, or require a huge amount of training data and time-consuming sequential decoders. In this paper, we propose an alternative paradigm. We model TSR as a logical location regression problem and propose a new TSR framework called LORE, standing for LOgical location REgression network, which for the first time combines logical location regression together with spatial location regression of table cells. Our proposed LORE is conceptually simpler, easier to train and more accurate than previous TSR models of other paradigms. Experiments on standard benchmarks demonstrate that LORE consistently outperforms prior arts. Code is available at https:// github.com/AlibabaResearch/AdvancedLiterateMachinery/tree/main/DocumentUnderstanding/LORE-TSR.

A Fault Diagnosis Method for Rotating Machinery With Semi-Supervised Graph Convolutional Network and Images Converted From Vibration Signals

In recent years, intelligent fault diagnosis has been widely investigated. Deep learning fault diagnosis methods for rotating machinery are primarily based on the assumption that labeled data are abundant. However, it is difficult to collect sufficient labeled samples for some practical engineering applications. To address this problem, a fault diagnosis method for rotating machinery with semi-supervised graph convolutional network (SSGCN) and two-dimensional (2-D) images converted from vibration signals is proposed in this paper. First, the one-dimensional (1-D) vibration signals are converted into 2-D images, by the no-parameter adjacency symmetrized dot pattern (NPASDP) method in polar coordinate system and the grayscale matrix mapping (GMM) method in Cartesian coordinate system. The local features of all samples are fully extracted using image processing methods in different coordinate systems to better characterize the relationship between different samples. Second, the normalized cross-correlation coefficient (NCC) is used to calculate the images similarity. Using each sample as a node, the graph structure is constructed by judging the adjacency relationship between all nodes through the similarity threshold. Finally, the SSGCN model is established to realize fault diagnosis. A bearing dataset and a rotor dataset with few data labels are used for validation, and the proposed method can get satisfactory fault diagnosis accuracy. The results show that the method can effectively improve the accuracy of model fault diagnosis with few labeled data.

Digital workflow for proximal adjustment of teeth adjacent to implant-supported fixed restorations using a predesigned computer-aided grinding guide.

Clinically, the proximal surfaces of teeth adjacent to an implant restoration usually need to be adjusted to build a preferable adjacency relationship. However, it is difficult for freehand preparation to get a favorable proximal contour in some cases. In the workflow presented here, virtual grinding can be made to adjacent teeth, under the consideration of functional reconstruction and biological requirements, and then the grinding can be implemented using digital templates and a specialized bur. This allows for more precise and accurate adjustments to be made during the clinical procedure, reducing the risk of over- or under-preparation of the proximal surfaces. In addition, the use of specialized diamond burs and grinding guides can make the procedure more efficient and streamlined, reducing the time required for proximal adjustment and minimizing patient discomfort. The resulting implant-supported prosthesis is more likely to function properly and last longer, as the precise proximal contacts can help distribute occlusal forces more evenly across the dentition. Overall, the use of digital technology for precise adjustment of proximal contacts during implant restorations represents an important advancement in modern dentistry, enabling dentists to provide their patients with more accurate, efficient, and effective dental care.

Leveraging multimodal features for knowledge graph entity alignment based on dynamic self-attention networks

Entity alignment establishes correspondence between entities across different knowledge graphs, which plays an important role in knowledge fusion. The key of entity alignment is to represent and fuse the features of entities across graphs. To improve efficiency, recent methods introduce graph embedding techniques that map different knowledge graphs into common latent spaces to encode structure and attribute features of entities. However, the structure embedding is generated by aggregating coarse connection information between entities without capturing semantics of the relations, which is incapable of dealing with entities of similar structures. Besides, multiple features are integrated by static arithmetic weighting method, which has difficulty in balancing the relative importance of structure and attribute features. To address these problems, we propose a dynamic self-attention based entity alignment model (DSEA), which leverages multimodal features for entity alignment through a dynamic self-attention network. To learn informative representations, we build a multimodal feature embedding component (MFE) to leverage structure, name, and position importance of entities. In MFE, semantics of adjacent entities and relations are exploited to enhance the structure embedding through a fine-granularity semantic augmentation network. Then a dynamic self-attention network is designed to adaptively evaluate the relative importance of structure and attribute features by learning the weights according to their contribution in entity alignment. Based on the feature weights, a similarity function of entities is formulated and the entity correspondence is established according to the similarities of entities. Extensive experiments conducted on five real-world datasets show that DSEA significantly outperforms state-of-the-art methods.

First look at the evaluation of three-loop non-planar Feynman diagrams for Higgs plus jet production

We present new computations for Feynman integrals relevant to Higgs plus jet production at three loops, including first results for a non-planar class of integrals. The results are expressed in terms of generalised polylogarithms up to transcendental weight six. We also provide the full canonical differential equations, which allows us to make structural observations on the answer. In particular, we find a counterexample to previously conjectured adjacency relations, for a planar integral of the tennis-court type. Additionally, for a non-planar triple ladder diagram, we find two novel alphabet letters. This information may be useful for future bootstrap approaches.

Automatic surface mesh generation by a binary-tree method

In order to realize the automation of surface mesh generation for arbitrary complex models, this paper proposes a binary-tree method (BTM) which can generate discontinuous grids. Compared with continuous grids, the adjacency relationship between discontinuous grids generated by BTM is more flexible, so it is easier to discretize “dirty” geometry without geometry repair than continuous grids. The BTM can adaptively generate meshes by features based on the curvature of curves and surfaces, and thus it does not need to construct a sizing field. Anisotropic meshes can also be directly generated by selecting the direction of subdividing meshes. Moreover, according to the periodic characteristics of periodic surfaces, the topological structure of the initial mesh can be designed to achieve mesh generation on periodic surfaces. Successful numerical examples are used to illustrate the effectiveness and robustness of the presented methods for meshing various types of geometric surfaces and reveal the excellent performance of the BTM. In addition, the BTM has been employed by the dual interpolation boundary face method (DiBFM) to solve 3D problems.

Enhancing context representations with part-of-speech information and neighboring signals for question classification

Question classification is an essential task in question answering (QA) systems. An effective and efficient question classification model can not only restrict the search space for answers, but also guide the QA system in selecting the optimal knowledge base and search strategy. In recent years, self-attention mechanism has been widely used in question classification for its strength of capturing global dependencies. However, it models all signals with weighted averaging, which is prone to overlooking the relation of neighboring signals. Furthermore, recent research has revealed that part-of-speech (POS) information can be used to determine and reinforce the semantics in sentence representation. In this paper, we propose a POS-aware adjacent relation attention network (POS-ARAN) for question classification, which enhance context representations with POS information and neighboring signals. To consider the local context, we propose an adjacent relation attention mechanism, which incorporates a Gaussian bias via a dynamic window to revise the vanilla self-attention mechanism. Thus, it can capture both the long-term dependency and local representation of semantic relations among words in different sentences. In addition, a POS-aware embedding layer is proposed, which helps to locate the appropriate headwords by syntactic information. Extensive experiments are conducted on Experimental Data for Question Classification (EDQC) dataset and Yahoo! Answers Comprehensive Questions and Answers 1.0, the results demonstrate that our model significantly outperforms the existing methods, achieving 95.59% in coarse-grained level accuracy and 92.91% in fine-grained level accuracy, respectively.

Video Process Mining and Model Matching for Intelligent Development: Conformance Checking.

Traditional business process-extraction models mainly rely on structured data such as logs, which are difficult to apply to unstructured data such as images and videos, making it impossible to perform process extractions in many data scenarios. Moreover, the generated process model lacks analysis consistency of the process model, resulting in a single understanding of the process model. To solve these two problems, a method of extracting process models from videos and analyzing the consistency of process models is proposed. Video data are widely used to capture the actual performance of business operations and are key sources of business data. Video data preprocessing, action placement and recognition, predetermined models, and conformance verification are all included in a method for extracting a process model from videos and analyzing the consistency between the process model and the predefined model. Finally, the similarity was calculated using graph edit distances and adjacency relationships (GED_NAR). The experimental results showed that the process model mined from the video was better in line with how the business was actually carried out than the process model derived from the noisy process logs.

Research on Remote Regional Collaborative Intelligence Platform Based on Blockchain

The 19th and 20th National Congresses of the Communist Party of China continued to stress the need to vigorously promote coordinated development among regions, making it an important strategic task in building a great modern socialist country in the new era. At present, the economic circle vigorously developed is based on the traditional regional cooperation based on the geographical adjacency relationship, while the regional cooperation obstacles will make the cooperation fail. Therefore, this paper builds the remote regional cooperation platform based on the virtual and real integration of game subject mirroring technology, collaborative contract execution technology of smart contract and collaborative credit management technology of blockchain, so as to realize the remote regional cooperation.

Adaptive Spatiotemporal InceptionNet for Traffic Flow Forecasting

Traffic flow forecasting is crucial to Intelligent Transportation Systems (ITS), particularly for route planning and traffic management. Spatiotemporal graph neural networks have been widely used for this purpose, where a global graph structure is mainly used, but local graph features are ignored. Multi-scale spatiotemporal features—the combined application of global and local graph features—can help effectively discover the underlying spatiotemporal patterns. This study proposed a multi-scale adaptive spatiotemporal forecasting model called Adaptive SpatioTemporal InceptionNet (AST-InceptionNet). To enrich the captured features and extract multi-scale spatiotemporal ones, the inception part is applied to combine local spatiotemporal features with several global ones. Additionally, we developed a fully adaptive graph convolution method, including topologically adaptive graph convolution and an adaptive adjacency matrix. It can autonomously and dynamically learn spatial heterogeneity and enable the proposed model to work with unknown adjacency relations. The proficiency of AST-InceptionNet was validated using four publicly available real-world traffic flow datasets. The experiments demonstrate a satisfactory performance from the proposed model, outperforming existing state-of-the-art methods by up to approximately 12%.

Ensemble-based Deep Learning Approach for Performance Improvement of BIM Element Classification

Ensemble learning is explored as an approach to improve the classification of BIM elements using multiple deep learning classifiers. Existing studies have focused on improving the performance of a single classifier, thereby restricting the types of features available for training. Two classifiers, MVCNN and MLP were trained on nine BIM element classes extracted from a 20-story office building. MVCNN used multiple images of these elements, while MLP used adjacency relations between elements. The predictions of the classifiers were then ensembled using a maximum rule-based algebraic combiner. Results showed that the overall classification accuracy improved by +0.05. The Ensemble model, in effect, leveraged the strength of each classifier: MVCNN for elements with distinct features and MLPs for elements with high similarity. The Ensemble model enables the automatic checking of the semantic accuracy of elements in BIM models, thus endorsing the seamless information exchanges required for true interoperability in BIM-centric project collaboration.