Graph Construction Research Articles

Scene-Driven Multimodal Knowledge Graph Construction for Embodied AI

Scene-Driven Multimodal Knowledge Graph Construction for Embodied AI

Leveraging recommendations using a multiplex graph database

PurposeBy applying targeted graph algorithms, the method used by the authors enables effective prediction of user interactions and thus fulfils the complex requirements of modern recommender systems. This study sets a new benchmark for multidimensional recommendation strategies and offers a path towards more advanced and user-centric models.Design/methodology/approachTo improve multidimensional data recommendation systems, multiplex graph structures are useful to capture various types of user interactions. This paper presents a novel framework that uses a graph database to compute and manipulate multiplex graphs. The approach enables flexible dimension management and increases expressive power through a specialised algebra designed for multiplex graph manipulation.FindingsThe authors compare the multiplex graph approach with traditional matrix methods, in particular random walk with restart, and show that the method not only provides deeper insights into user preferences by integrating scores from different layers of the multiplex graph, but also outperforming matrix-based approaches in most configurations. The results highlight the potential of multiplex graphs for developing sophisticated and customised recommender systems that significantly improve both performance and explainability.Originality/valueThe study provides a formal specification of a multiplex graph construction based on interaction and content-based information; and the study also developed an algebra dedicated to multiplex graphs, enabling robust and precise graph manipulations necessary for effective recommendation queries. The authors implement these algebraic operations within the Neo4j graph database system with a thorough analysis and experimentation with three different data sets, benchmarked against traditional matrix-based methods.

Ensemble-Enhanced Semi-Supervised Learning with Optimized Graph Construction for High-Dimensional Data.

Graph-based methods have demonstrated exceptional performance in semi-supervised classification. However, existing graph-based methods typically construct either a predefined graph in the original space or an adaptive graph within the output space, which often limits their ability to fully utilize prior information and capture the optimal intrinsic data distribution, particularly in high-dimensional data with abundant redundant and noisy features. This paper introduces a novel approach: Semi-Supervised Classification with Optimized Graph Construction (SSC-OGC). SSC-OGC leverages both predefined and adaptive graphs to explore intrinsic data distribution and effectively employ prior information. Additionally, a graph constraint regularization term (GCR) and a collaborative constraint regularization term (CCR) are incorporated to further enhance the quality of the adaptive graph structure and the learned subspace, respectively. To eliminate the negative effect of constructing a predefined graph in the original data space, we further propose a Hybrid Subspace Ensemble-enhanced framework based on the proposed Optimized Graph Construction method (HSE-OGC). Specifically, we construct multiple hybrid subspaces, which consist of meticulously chosen features from the original data to achieve high-quality and diverse space representations. Then, HSE-OGC constructs multiple predefined graphs within hybrid subspaces and trains multiple SSC-OGC classifiers to complement each other, significantly improving the overall performance. Experimental results conducted on various high-dimensional datasets demonstrate that HSE-OGC exhibits outstanding performance.

Construction of Legal Knowledge Graph Based on Knowledge-Enhanced Large Language Models

Legal knowledge involves multidimensional heterogeneous knowledge such as legal provisions, judicial interpretations, judicial cases, and defenses, which requires extremely high relevance and accuracy of knowledge. Meanwhile, the construction of a legal knowledge reasoning system also faces challenges in obtaining, processing, and sharing multisource heterogeneous knowledge. The knowledge graph technology, which is a knowledge organization form with triples as the basic unit, is able to efficiently transform multisource heterogeneous information into a knowledge representation form close to human cognition. Taking the automated construction of the Chinese legal knowledge graph (CLKG) as a case scenario, this paper presents a joint knowledge enhancement model (JKEM), where prior knowledge is embedded into a large language model (LLM), and the LLM is fine-tuned through the prefix of the prior knowledge data. Under the condition of freezing most parameters of the LLM, this fine-tuning scheme adds continuous deep prompts as prefix tokens to the input sequences of different layers, which can significantly improve the accuracy of knowledge extraction. The results show that the knowledge extraction accuracy of the JKEM in this paper reaches 90.92%. Based on the superior performance of this model, the CLKG is further constructed, which contains 3480 knowledge triples composed of 9 entities and 2 relationships, providing strong support for an in-depth understanding of the complex relationships in the legal field.

Time-Efficient RSA over Large-Scale Multi-Domain EON

The poor timeliness of routing has always been an urgent problem in practical operator networks, especially in situations with large-scale networks and multiple network domains. In this article, a pruning idea of routing integrated with Dijkstra’s shortest path searching is utilized to accelerate the process of routing in large-scale multi-domain elastic optical networks (EONs). The layered-graph approach is adopted in the spectrum allocation stage. To this end, an efficient heuristic algorithm is proposed, called “Branch-and-Bound based Routing and Layered Graph based Spectrum Allocation algorithm (BBR-LGSA)”, which is an integrated RSA algorithm. Notably, the significant reduction in algorithm time complexity is not only reflected in the pruning method used in the routing stage but also in the construction of auxiliary graphs during the spectrum allocation stage utilizing the Branch-and-Bound method. Simulation results show that the proposed BBR-LGSA significantly reduces the average running time by nearly 78% with higher spectrum utilization in large-scale multi-domain EONs, compared with benchmark algorithms. In addition, the impact of key parameters on performance comparisons of different algorithms is evaluated.

Whole slide cervical cancer classification via graph attention networks and contrastive learning

Cervical cancer is one of the most common cancers among women, which seriously threatens women’s health. Early screening can reduce the incidence rate and mortality. Thinprep cytologic test (TCT) is one of the important means of cytological screening, which has high sensitivity and specificity, and has been widely used in the early screening of cervical cancer. The automatic diagnosis of whole slide images (WSIs) by computers can effectively improve the efficiency and accuracy of doctors’ diagnoses. However, current methods ignore the intrinsic relationships between cervical cells in WSIs and neglect contextual information from the surrounding suspicious areas, and therefore limit their robustness and generalizability. In this paper, we propose a novel two-stage method to implement the automatic diagnosis of WSIs, which constructs Graph Attention Networks (GAT) based on local and global fields respectively to capture their contextual information in a hierarchical manner. In the first stage, we extract representative patches from each WSI through suspicious cell detection, and then employ a Local GAT to classify cervical cells by capturing correlations between suspicious cells in image tiles. This classification process provides the confidence and feature vectors for each suspicious cell. In the second stage, we perform WSI classification using a Global GAT model. We construct graphs corresponding to top-Kg and bottom-Kg cells for each WSI based on results from Local GAT, and introduce a supervised contrastive learning strategy to enhance the discriminative power of the extracted features. Experimental results demonstrate that our proposed method outperforms conventional approaches and effectively showcases the benefits of supervised contrastive learning. Our source code and example data are available at https://github.com/feimanman/Whole-Slide-Cervical-Cancer-Classification.

A knowledge graph based remanufacturing equipment resource modeling method.

Background The uncertainty and non-standardization of the remanufacturing process result in multi-source heterogeneity of remanufacturing equipment capabilities and knowledge information, which limits the development of informationization and intelligence in remanufacturing. Methods Therefore, we propose a knowledge graph-based modeling method for Remanufacturing Equipment Resource (RMER) information. By analyzing the categories of remanufacturing equipment resources and integrating resource information such as man, machine, material, method, and environment, a unified remanufacturing equipment information model and service capability information model were constructed. Based on the top-down knowledge graph construction process, the RMER semantic association information of the ontology pattern and data layers was extracted and integrated to construct the RMER knowledge graph. Results In this paper, 18 enterprise texts, 50 experience texts and thesis patent texts were selected as the training set, and 2 enterprise texts, 50 experience texts and thesis patent texts were selected as the test set. After multiple rounds of model training, the accuracy rate of entity information extraction reached 91.4%, which proved the accuracy and effectiveness of the model for entity information document extraction. Conclusions This study provides a unified information foundation for mapping and optimizing the resource capabilities of remanufacturing equipment. It can help remanufacturing enterprises to quickly configure equipment resources, and improve the intelligent and automated management level of RMER.

Unsafe behavior identification on construction sites by combining computer vision and knowledge graph–based reasoning

PurposeWorkers' unsafe behavior is the main cause of construction safety accidents, thereby highlighting the critical importance of behavior-based management. To compensate for the limitations of computer vision in tackling knowledge-intensive issues, semantic-based methods have gained increasing attention in the field of construction safety management. Knowledge graph provides an efficient and visualized method for the identification of various unsafe behaviors.Design/methodology/approachThis study proposes an unsafe behavior identification framework by integrating computer vision and knowledge graph–based reasoning. An enhanced ontology model anchors our framework, with image features from YOLOv5, COCO Panoptic Segmentation and DeepSORT integrated into the graph database, culminating in a structured knowledge graph. An inference module is also developed, enabling automated the extraction of unsafe behavior knowledge through rule-based reasoning.FindingsA case application is implemented to demonstrate the feasibility and effectiveness of the proposed method. Results show that the method can identify various unsafe behaviors from images of construction sites and provide mitigation recommendations for safety managers by automated reasoning, thus supporting on-site safety management and safety education.Originality/valueExisting studies focus on spatial relationships, often neglecting the diversified spatiotemporal information in images. Besides, previous research in construction safety only partially automated knowledge graph construction and reasoning processes. In contrast, this study constructs an enhanced knowledge graph integrating static and dynamic data, coupled with an inference module for fully automated knowledge-based unsafe behavior identification. It can help managers grasp the workers’ behavior dynamics and timely implement measures to correct violations.

Cluster-efficient pangenome graph construction with nf-core/pangenome.

Pangenome graphs offer a comprehensive way of capturing genomic variability across multiple genomes. However, current construction methods often introduce biases, excluding complex sequences or relying on references. The PanGenome Graph Builder (PGGB) addresses these issues. To date, though, there is no state-of-the-art pipeline allowing for easy deployment, efficient and dynamic use of available resources, and scalable usage at the same time. To overcome these limitations, we present nf-core/pangenome, a reference-unbiased approach implemented in Nextflow following nf-core's best practices. Leveraging biocontainers ensures portability and seamless deployment in HPC environments. Unlike PGGB, nf-core/pangenome distributes alignments across cluster nodes, enabling scalability. Demonstrating its efficiency, we constructed pangenome graphs for 1000 human chromosome 19 haplotypes and 2146 E. coli sequences, achieving a two to threefold speedup compared to PGGB without increasing greenhouse gas emissions. Nf-core/pangenome is released under the MIT open-source license, available on GitHub and Zenodo, with documentation accessible at https://nf-co.re/pangenome/1.1.2/docs/usage. Supplementary data are available at Bioinformatics online.

A Novel Kind of Knowledge Graph Construction Method for Intelligent Machine as a Service Modeling

With the development of Intelligent Machine as a Service (IMaaS), devices increasingly require personalization, intelligence, and service orientation, making resource modeling a key challenge. Knowledge graph (KG) technology, known for unifying heterogeneous data, has become an essential tool for modeling and analyzing manufacturing resources. On this basis, this study proposes a novel resource KG construction method for IMaaS. First, an E-R diagram is used to divide the constant and variable entities and set the constant attributes and the constant relationships. Then, the triplets are named, the value space is set, and the schema layer is constructed. Finally, the related information about devices is used to fill the data layer, and then, the knowledge graph is generated. Meanwhile, this study utilizes desktop FDM 3D printing devices as a case example for validation. The method proposed in this study can enhance the accuracy and maintainability of equipment resource management in the manufacturing sector, effectively promoting subsequent activities such as management, analysis, and decision-making.

Research on the Construction Method of Curriculum Teaching Knowledge Graph Based on Bi-LSTM and CNN Algorithm

The aim of the paper is to explore a method of constructing a curriculum teaching knowledge graph by combining Bi-LSTM and convolutional neural network (CNN) algorithm. The field of education is constantly seeking innovation to improve teaching results and student learning experience. Knowledge graph, as an advanced technology of structured representation of knowledge, is expected to provide effective support for teaching management and personalized learning. First, this paper introduces the background and significance of the curriculum teaching knowledge graph. By establishing knowledge maps, we can more clearly present the knowledge system and correlation in the curriculum, which helps teachers to design more targeted teaching content and provide personalized learning paths for students. However, traditional knowledge graph construction methods are often faced with problems such as incomplete information capture and inaccurate semantic association, so it is necessary to introduce advanced deep learning algorithms to improve the quality of knowledge graph. Secondly, this paper elaborates on the construction method of fusion Bi-LSTM and CNN algorithm. Bi-LSTM, as a recurrent neural network capable of capturing sequence information, can better model the evolution process of knowledge in the course. As a CNN is good at extracting local features, CNN can effectively capture the spatial structure information in the knowledge graph. By integrating two, we can improve the expression ability and reasoning accuracy of knowledge graph. Further, the experimental results show that the fusion Bi-LSTM and the CNN algorithm have significantly improved the accuracy of information capture and inference compared with the traditional method. In summary, this paper proposes an innovative construction method of curriculum teaching knowledge graph by integrating Bi-LSTM and CNN algorithm, which provides new ideas and solutions for informatization and personalized teaching in the field of education. In the future, the applicability of this method in different disciplines and teaching scenarios can be further discussed, and more advanced technologies can be combined to continuously improve and expand the research.

Enhanced tensor based embedding anchor learning for multi-view clustering

Existing anchor graph based multi-view clustering methods can overcome the problem of high computational cost in traditional multi-view clustering methods. However, the anchor points selected from high-dimensional data often contain irrelevant noise and outliers, which would affect the clustering performance. To address this issue, we propose an embedding anchor based multi-view clustering method, called enhanced tensor based embedding anchor learning (ETEAL). Specifically, we unify the learning process of latent embedding space, anchor points, and anchor graphs into a common framework, which eliminates noise and redundant data in the original space and enhances the synergistic optimization between the components. Meanwhile, we employ an enhanced tensor strategy to constrain the embedding anchor graphs, which exploits the higher-order relationships between views and recovers the global low-rank property of the embedding anchor graphs. Finally, we develop an anchor graph fusion strategy, which significantly reduces the huge overhead of traditional graph fusion that requires the construction of complete graphs. Experimental results on eight benchmark datasets show that the proposed method significantly outperforms other state-of-the-art methods in terms of scalability and clustering accuracy.

Marginal Effect-aware Multiple-Vehicle Scheduling for Road Data Collection: A Near-optimal Result

Vehicles equipped with abundant sensors offer a promising way for large-scale, low-cost road data collection. To realize this potential, a well-designed vehicle scheduling scheme is essential for deploying the recruited drivers efficiently. Nevertheless, existing works fail to consider the marginal effect among drivers’ collections. Different from them, this article introduces a, to the best of our knowledge, new multiple-vehicle scheduling problem that jointly optimizes task allocation and vehicle trajectory planning to maximize the overall collection utility by accounting for the marginal effect in drivers’ data collections. However, solving this problem is non-trivial due to its involvement with multiple coupled NP-hard problems. To this end, we propose MeSched, a marginal effect-aware multiple-vehicle scheduling scheme designed for road data collection. Specifically, we first present a greedy-based auxiliary graph construction method to disentangle the initial problem into multiple independent single-vehicle scheduling subproblems. Furthermore, we build an approximate surrogate function that transforms each subproblem into a tractable form involving only a single variable. The theoretical analysis proves that MeSched can achieve a 1-(1/ e ) ¼ -approximation ratio in polynomial time. Comprehensive evaluations based on a real-world trajectory dataset of 12,493 vehicles demonstrate that MeSched can significantly improve the collection utility by 104.5% on average compared with four baselines.

Information Extraction of Aviation Accident Causation Knowledge Graph: An LLM-Based Approach

Summarizing the causation of aviation accidents is conducive to enhancing aviation safety. The knowledge graph of aviation accident causation, constructed based on aviation accident reports, can assist in analyzing the causes of aviation accidents. With the continuous development of artificial intelligence technology, leveraging large language models for information extraction and knowledge graph construction has demonstrated significant advantages. This paper proposes an information extraction method for aviation accident causation based on Claude-prompt, which relies on the large-scale pre-trained language model Claude 3.5. Through prompt engineering, combined with a few-shot learning strategy and a self-judgment mechanism, this method achieves automatic extraction of accident-cause entities and their relationships. Experimental results indicate that this approach effectively improves the accuracy of information extraction, overcoming the limitations of traditional methods in terms of accuracy and efficiency in processing complex texts. It provides strong support for subsequently constructing a structured knowledge graph of aviation accident causation and conducting causation analysis of aviation accidents.

A novel label-aware global graph construction method and spiking-coded graph neural network for intelligent process fault diagnosis

Fault diagnosis plays a crucial role in ensuring the safety and efficiency of industrial processes. However, traditional techniques often face difficulties in handling large-scale data characterized by complex structures and relationships. To efficiently represent industrial data as graphs and develop a low-energy-cost feature extraction model, a novel label-aware global graph construction method and a spiking graph convolutional network (SGCN) are proposed in this study to achieve intelligent process fault diagnosis. The label-aware method enhances graph data representation by capturing intrinsic correlations and global features. The SGCN integrates graph convolutional layers with spiking encoding, enabling effective feature extraction while offering computational efficiency advantages. A weighted loss function is introduced to mitigate data imbalance issues. Experiments on the Tennessee Eastman process, the Three-phase Flow Facility, and the de-propanizer distillation process demonstrate SGCN’s superior performance over baseline models in various fault scenarios, while significantly reducing computational costs. The proposed method offers promising potential for reliable and efficient fault diagnosis in complex real-world industrial environments.

BioSAM: Generating SAM Prompts From Superpixel Graph for Biological Instance Segmentation.

Proposal-free instance segmentation methods have significantly advanced the field of biological image analysis. Recently, the Segment Anything Model (SAM) has shown an extraordinary ability to handle challenging instance boundaries. However, directly applying SAM to biological images that contain instances with complex morphologies and dense distributions fails to yield satisfactory results. In this work, we propose BioSAM, a new biological instance segmentation framework generating SAM prompts from a superpixel graph. Specifically, to avoid over-merging, we first generate sufficient superpixels as graph nodes and construct an initialized graph. We then generate initial prompts from each superpixel and aggregate them through a graph neural network (GNN) by predicting the relationship of superpixels to avoid over-segmentation. We employ the SAM encoder embeddings and the SAM-assisted superpixel similarity as new features for the graph to enhance its discrimination capability. With the graph-based prompt aggregation, we utilize the aggregated prompts in SAM to refine the segmentation and generate more accurate instance boundaries. Comprehensive experiments on four representative biological datasets demonstrate that our proposed method outperforms state-of-the-art methods.

A unified active learning framework for annotating graph data for regression tasks

In many domains, effectively applying machine learning models requires a large number of annotations and labelled data, which might not be available in advance. Acquiring annotations often requires significant time, effort, and computational resources, making it challenging. Active learning strategies are pivotal in addressing these challenges, particularly for diverse data types such as graphs. Although active learning has been extensively explored for node-level classification, its application to graph-level learning, especially for regression tasks, is not well-explored. We develop a unified active learning framework specializing in graph annotating and graph-level learning for regression tasks on both standard and expanded graphs, which are more detailed representations. We begin with graph collection and construction. Then, we construct various graph embeddings (unsupervised and supervised) into a latent space. Given such an embedding, the framework becomes task agnostic and active learning can be performed using any regression method and query strategy suited for regression. Within this framework, we investigate the impact of using different levels of information for active and passive learning, e.g., partially available labels and unlabelled test data. Despite our framework being domain agnostic, we validate it on a real-world application of software performance prediction, where the execution time of the source code is predicted. Thus, the graph is constructed as an intermediate source code representation. We support our methodology with a real-world dataset to underscore the applicability of our approach. Our real-world experiments reveal that satisfactory performance can be achieved by querying labels for only a small subset of all the data. A key finding is that Graph2Vec (an unsupervised embedding approach for graph data) performs the best, but only when all train and test features are used. However, Graph Neural Networks (GNNs) are the most flexible embedding techniques when used for different levels of information with and without label access. In addition, we find that the benefit of active learning increases for larger datasets (more graphs) and when the graphs are more complex, which is arguably when active learning is the most important.

Ship model-based route optimisation for decision support in deep sea shipping

Abstract We present a new approach and route optimization methodology to support analysis and planning of vessel and fleet performance in deep sea shipping for green, energy efficient, and safe navigation. The developed methodology combines the ship technical characteristics based on a ship model developed for a particular vessel type and energy-saving technology options and an optimization algorithm taking into account weather conditions. Optimization involves two stages: graph construction and Dynamic Programming labelling algorithm implemented to solve the shortest path problem with variable speed. The new approach is implemented as part of a decision-support tool EcoRouter enabling the user to conduct analysis of safe and Pareto optimal solutions. Several applications to support real fleet planning and ship performance analysis have been identified including project engineering for future energy-saving ship technologies, for example, wind-assisted propulsion.

Taking relations as known conditions: A tagging based method for relational triple extraction

Relational triple extraction refers to extracting entities and relations from natural texts, which is a crucial task in the construction of knowledge graph. Recently, tagging based methods have received increasing attention because of their simple and effective structural form. Among them, the two-step extraction method is easy to cause the problem of category imbalance. To address this issue, we propose a novel two-step extraction method, which first extracts subjects, generates a fixed-size embedding for each relation, and then regards these relations as known conditions to extract the objects directly with the identified subjects. In order to eliminate the influence of irrelevant relations when predicting objects, we use a relation-special attention mechanism and a gate unit to select appropriate relations. In addition, most current models do not account for two-way interaction between tasks, so we design a feature interactive network to achieve bidirectional interaction between subject and object extraction tasks and enhance their connection. Experimental results on NYT, WebNLG, NYT⋆ and WebNLG⋆ datasets show that our model is competitive among joint extraction models.

PLRTE: Progressive learning for biomedical relation triplet extraction using large language models

Document-level relation triplet extraction is crucial in biomedical text mining, aiding in drug discovery and the construction of biomedical knowledge graphs. Current language models face challenges in generalizing to unseen datasets and relation types in biomedical relation triplet extraction, which limits their effectiveness in these crucial tasks. To address this challenge, our study optimizes models from two critical dimensions: data-task relevance and granularity of relations, aiming to enhance their generalization capabilities significantly. We introduce a novel progressive learning strategy to obtain the PLRTE model. This strategy not only enhances the model’s capability to comprehend diverse relation types in the biomedical domain but also implements a structured four-level progressive learning process through semantic relation augmentation, compositional instruction and dual-axis level learning. Our experiments on the DDI and BC5CDR document-level biomedical relation triplet datasets demonstrate a significant performance improvement of 5% to 20% over the current state-of-the-art baselines. Furthermore, our model exhibits exceptional generalization capabilities on the unseen Chemprot and GDA datasets, further validating the effectiveness of optimizing data-task association and relation granularity for enhancing model generalizability.