Attributed Graphs Research Articles

A Biomedical Open Knowledge Network Harnesses the Power of AI to Understand Deep Human Biology

Knowledge representation and reasoning (KR&R) has been successfully implemented in many fields to enable computers to solve complex problems with AI methods. However, its application to biomedicine has been lagging in part due to the daunting complexity of molecular and cellular pathways that govern human physiology and pathology. In this article, we describe concrete uses of Scalable PrecisiOn Medicine Knowledge Engine (SPOKE), an open knowledge network that connects curated information from thirty-seven specialized and human-curated databases into a single property graph, with 3 million nodes and 15 million edges to date. Applications discussed in this article include drug discovery, COVID-19 research and chronic disease diagnosis, and management.

Context mining and graph queries on giant biomedical knowledge graphs

Contextual information is widely considered for NLP and knowledge discovery in life sciences since it highly influences the exact meaning of natural language. The scientific challenge is not only to extract such context data, but also to store this data for further query and discovery approaches. Classical approaches use RDF triple stores, which have serious limitations. Here, we propose a multiple step knowledge graph approach using labeled property graphs based on polyglot persistence systems to utilize context data for context mining, graph queries, knowledge discovery and extraction. We introduce the graph-theoretic foundation for a general context concept within semantic networks and show a proof of concept based on biomedical literature and text mining. Our test system contains a knowledge graph derived from the entirety of PubMed and SCAIView data and is enriched with text mining data and domain-specific language data using Biological Expression Language. Here, context is a more general concept than annotations. This dense graph has more than 71M nodes and 850M relationships. We discuss the impact of this novel approach with 27 real-world use cases represented by graph queries. Storing and querying a giant knowledge graph as a labeled property graph is still a technological challenge. Here, we demonstrate how our data model is able to support the understanding and interpretation of biomedical data. We present several real-world use cases that utilize our massive, generated knowledge graph derived from PubMed data and enriched with additional contextual data. Finally, we show a working example in context of biologically relevant information using SCAIView.

PG-explorer: Resource Description Framework data exploration with property graphs

The Resource Description Framework (RDF) has been widely applied to represent and exchange domain information because of its machine-readable characteristic. With a huge amount of RDF data available, retrieving RDF data is essential, so that many RDF query approaches have been developed. But many traditional approaches generally require users to know RDF model and query language, and this seriously prevents a large number of common non-expert users from obtaining information in RDF datasets. In this paper, we propose an approach that users can explore massive RDF datasets by interactively manipulating property graphs. Our approach provides users with a series of operations in interactively constructing property graphs to describe their query intents. To efficiently explore massive RDF datasets, we convert RDF data into property graphs for storage. This can greatly reduce the size of massive RDF datasets and more importantly the converted property graphs can instruct users to understand the underlying structure of RDF datasets, which is very useful in users’ construction of their query property graphs. The constructed query property graphs are finally transformed into expressions with the query language of graph databases. With high-performance graph databases as query engines, we developed an RDF data exploration system – PG-Explorer. Through experiments over real-world datasets, we evaluated the effectiveness and superiority of our approach. The user study demonstrates that the proposed approach simplifies users’ exploration of RDF data and can satisfy their exploration needs.

A biomedical open knowledge network harnesses the power of AI to understand deep human biology.

Knowledge representation and reasoning (KR&R) has been successfully implemented in many fields to enable computers to solve complex problems with AI methods. However, its application to biomedicine has been lagging in part due to the daunting complexity of molecular and cellular pathways that govern human physiology and pathology. In this article we describe concrete uses of SPOKE, an open knowledge network that connects curated information from 37 specialized and human-curated databases into a single property graph, with 3 million nodes and 15 million edges to date. Applications discussed in this article include drug discovery, COVID-19 research and chronic disease diagnosis and management.

A Generic Graph-Based Method for Flexible Aspect-Opinion Analysis of Complex Product Customer Feedback

Product design experts depend on online customer reviews as a source of insight to improve product design. Previous works used aspect-based sentiment analysis to extract insight from product reviews. However, their approaches for requirements elicitation are less flexible than traditional tools such as interviews and surveys. They require costly data labeling or pre-labeled datasets, lack domain knowledge integration, and focus more on sentiment classification than flexible aspect-opinion analysis. Related works lack effective mechanisms for probing the customer feedback of complex configurable products. This study proposes a generic graph-based opinion mining and analysis method for product design improvement. First, a customer feedback data preprocessing and annotation pipeline that can incorporate designer-specified domain knowledge is proposed. Second, an intuitive opinion-aware labeled property graph data model is designed to ingest preprocessed feedback data and perform ad hoc opinion analysis. Applying the generic model to a real-world dataset demonstrates superior functionality and flexibility compared to related works. A wider range of analyses is supported in a single model without repeating data preprocessing and modeling. Specifically, the proposed method supports regular and comparative aspect-opinion analysis, aspect satisfaction/influence ranking, opinion trend extraction, and targeted aspect-opinion summarization.

A unified Link Prediction architecture applied on a novel heterogenous Knowledge Base

Link Prediction (LP) aims at addressing incompleteness of Knowledge Graph (KG). The goal of LP is to capture the distribution of entities and relations present in a KG and utilise these to predict probability of missing information. State-of-the-art LP approaches rely on latent feature models for this purpose. The research focus has predominantly been on application of LP to triple based datasets (e.g. Freebase, YAGO). However, with growing adoption of KGs, it is common to see more heterogeneous property graphs being used, examples of common properties are temporal and weight data. The contributions of the following work are two fold. First, we introduce a novel framework which is the first to provide support for latent feature model LP on heterogeneous Knowledge Bases (KBs). Second, we utilise a novel KB — Refinitiv Knowledge Graph, to produce a heterogeneous dataset with which capabilities of the framework are examined.

Method of Domain Knowledge Graph Construction Based on Property Graph Model

Method of Domain Knowledge Graph Construction Based on Property Graph Model

A Modified Evolutionary Algorithm for Generating the Cycling Training Routes

We present a novel method for generating cycling training routes from geographical property graphs based on an Evolutionary Algorithm. The algorithm operators of crossover and mutation are adjusted for use in the Property Graph domain. Data fusion of geographical data from the OpenStreetMap, EU-DEM digital surface model, and existing training records is performed as a basis of the intersections-paths’ property graph. The proposed approach allows route generation based on their starting and ending points in the property graph and their distance and ascent. A property graph of all intersections and cycling roads is shown and generated for the regions of Podravje and Pomurje. The property graph used in the proposed algorithm’s feasibility demonstration is shown. This is done by presenting four different cases of routes that were generated with our algorithm. The algorithm allows for generating classic cycling routes of A to B nature, routes with more than two fixed points, as well as cyclical training routes. The research is concluded by offering further directions on route generation research.

Learning Content Classification and Mapping Content to Synonymous Learners based on 2022 Augmented Verb List of Marzano and Kendall Taxonomy

Finding suitable learning content for learners with different learning styles is a challenging task in the learning process. Hence it is essential to follow some learning taxonomies to deliver learner-centric learner content. Learning taxonomies are used to express various learning practices and learning habits to be followed by the learner for a better learning process. The investigator has already classified the learners based on the 2022 augmented verb list of Marzano and Kendall taxonomy. The main objective of this paper is to minutely classify the tutor-defined learning contents according to the domains as well as the subdomains of the considered taxonomy which is in text format. Providing personalized learning content could help the learners for a better understanding of learning content and their interrelationship which in turn produce better learning outcomes. Mapping the six levels of learning contents into the corresponding learner is a challenging task. Hence the investigator has chosen seven algorithms including Bagging, XG Boost, Support Vector Machine from Machine Learning and four algorithms including Convolutional Neural Network, and Deep Neural Network in Deep Learning algorithm to classify the learning contents. The experimental results indicate that Support Vector Machine performed well in machine learning and Deep Neural Network yields good performance in deep learning in the learning content classification process. These micro contents were organized using a property graph. Further, the micro contents were retrieved from the property graph using SPARQL for mapping the classified contents to the corresponding learners to achieve personalization in the learning process.

A General Cardinality Estimation Framework for Subgraph Matching in Property Graphs

We introduce a framework for cardinality estimation of query patterns over property graph databases. This framework makes it possible to analyze, compare and combine different cardinality estimation approaches. It consists of three phases: obtaining a set of estimates for some subqueries, extending this set and finally combining the set into a single cardinality estimate for the query. We show that (parts of) many existing cardinality estimation approaches can be used as techniques in one of the phases from our framework. The phases are loosely coupled, making it possible to combine (parts of) current cardinality estimation approaches. We created a graph version of the Join Order Benchmark to perform experiments with different combinations of techniques. The results showed that query patterns without property constraints can be accurately estimated using synopses for small patterns. Accurate estimation of query patterns with property constraints require new estimation techniques to be developed that capture correlations between the property constraints and the topology in graph databases.

Extending the Intelligent Digital Twin with a context modeling service: A decision support use case

Considering the surrounding context of manufacturing systems at runtime enhances their adaptability, which is advantageous given the increasing demand for flexible production. Context-awareness presents a viable approach to continuously model the surrounding context and relate it to the system’s operation. As the surrounding context is both acquirable from heterogeneous data sources and is dynamic in nature, a coherent and evolving context model is necessary. The context modeling approach needs to be both: system-centric to enable re-usability of the model by different applications as well as continuously and efficiently extendable to continuously represent the environment. Some approaches to model context within the manufacturing domain exist, but are mostly tailored to specific applications or do not consider an extension at runtime. To include the surrounding context of a system effectively and adapt it accordingly, a relation between the modeled context and the system during runtime is necessary. The intelligent Digital Twin provides essential interdisciplinary models of a physical system, which can be used for its analysis and monitoring. Adding a further service in the form of a model to represent the surrounding context enhances the intelligent Digital Twin to be used by various applications, e.g., for reconfiguration or decision-support. In our previous work, we have presented a tier-based model to realize context-awareness for intelligent Digital Twins. The presented tier model has different levels corresponding to different context scopes. In this contribution, we present an approach to realize an extendible and evolving tier-based context model with a generic ontology and model transformation to a labeled property graph with focus on the external context. To show the applicability and added value of the context model, a use case is presented, which addresses an intelligent warehouse’s Digital Twin and its context-model supported decision-making.

Graph Database to Enhance Supply Chain Resilience for Industry 4.0

Supply chain network in the automotive industry has complex, interconnected, multiple-depth relationships. Recently, the volume of supply chain data increases significantly with Industry 4.0. The complex relationships and massive volume of supply chain data can cause visibility and scalability issues in big data analysis and result in less responsive and fragile inventory management. The authors develop a graph data modeling framework to address the computational problem of big supply chain data analysis. In addition, this paper introduces Time-to-Stockout analysis for supply chain resilience and shows how to compute it through a labeled property graph model. The computational result shows that the proposed graph data model is efficient for recursive and variable-length data in supply chain, and relationship-centric graph query language has capable of handling a wide range of business questions with impressive query time.

COMPUTING WATERTIGHT VOLUMETRIC MODELS FROM BOUNDARY REPRESENTATIONS TO ENSURE CONSISTENT TOPOLOGICAL OPERATIONS

Abstract. To simulate environmental processes, noise, flooding in cities as well as the behaviour of buildings and infrastructure, ‘watertight’ volumetric models are a measuring prerequisite. They ensure topologically consistent 3D models and allow the definition of proper topological operations. However, in many existing city or other geo-information models, topologically unchecked boundary representations are used to store spatial entities. In order to obtain consistent topological models, including their ‘fillings’, in this paper, a triangulation combined with overlay and path-finding methods is presented by climbing up the dimension, beginning with the wireframe model. The algorithms developed for this task are presented, whereby using the philosophy of graph databases and the Property Graph Model. Examples to illustrate the algorithms are given, and experiments are performed on a data-set from Erfurt, Thuringia (Germany), providing complex geometries of buildings. The heavy influence of double precision arithmetic on the results, in particular the positional and angular precision, is discussed in the end.

Explainable zero-shot learning via attentive graph convolutional network and knowledge graphs

Zero-shot learning (ZSL) which aims to deal with new classes that have never appeared in the training data (i.e., unseen classes) has attracted massive research interests recently. Transferring of deep features learned from training classes (i.e., seen classes) are often used, but most current methods are black-box models without any explanations, especially textual explanations that are more acceptable to not only machine learning specialists but also common people without artificial intelligence expertise. In this paper, we focus on explainable ZSL, and present a knowledge graph (KG) based framework that can explain the transferability of features in ZSL in a human understandable manner. The framework has two modules: an attentive ZSL learner and an explanation generator. The former utilizes an Attentive Graph Convolutional Network (AGCN) to match class knowledge from WordNet with deep features learned from CNNs (i.e., encode inter-class relationship to predict classifiers), in which the features of unseen classes are transferred from seen classes to predict the samples of unseen classes, with impressive (important) seen classes detected, while the latter generates human understandable explanations for the transferability of features with class knowledge that are enriched by external KGs, including a domain-specific Attribute Graph and DBpedia. We evaluate our method on two benchmarks of animal recognition. Augmented by class knowledge from KGs, our framework generates promising explanations for the transferability of features, and at the same time improves the recognition accuracy.

Exploring the application of property graph model in visualizing COBie data

Purpose The usability aspect of the construction operations building information exchange (COBie) datasheet has been largely overlooked. Users find it difficult to find relevant data inside COBie and understand the dependencies of information. This research study is a part of a more comprehensive research study to identify the usability issues associated with COBie and propose solutions to deal with them. This paper aims to discuss the challenges associated with the visualization aspect of COBie and proposes a solution to mitigate them. Design/methodology/approach This paper is based on design thinking and waterfall methodology. While the design thinking methodology is used to explore the issues associated with the visualization aspect of COBie, the waterfall methodology is used to develop a working prototype of the visualizer for the COBie datasheet using a spreadsheet format. Findings The paper demonstrates that the property graph model based on a node-link diagram can be effectively used to represent the COBie datasheet. This will help in storing data in a visually connected manner and looking at links more dynamically. Moreover, converting and storing data into an appropriate database will help reach data directly rather than navigate multiple workbooks. This database can also help get the history of data inside the COBie datasheet as it develops throughout the project. Originality/value This research proposes a novel approach to visualize the COBie datasheet interactively using the property graph model, a type of node-link diagram. Using the property graph model will help users see data in a connected way, which is currently missing in the spreadsheet representation of COBie data. Moreover, this research also highlights that storing historical changes in COBie data can help understand how data has evolved throughout the construction. Additionally, structured storage of data in relationship format can help users to access the end of connected data directly through the efficient search.

Designing and Building Enterprise Knowledge Graphs

This book is a guide to designing and building knowledge graphs from enterprise relational databases in practice. It presents a principled framework centered on mapping patterns to connect relational databases with knowledge graphs, the roles within an organization responsible for the knowledge graph, and the process that combines data and people. The content of this book is applicable to knowledge graphs being built either with property graph or RDF graph technologies.

Constructing Crop Portraits Based on Graph Databases Is Essential to Agricultural Data Mining

Neo4j is a graph database that can use not only data, but also data relationships. Crop portraits, a kind of property graph, model the crop entity in the real world based on data to realize the networked management of crop knowledge. The existing crop knowledge base has shortcomings such as single crop variety, incomplete description, and lack of agricultural knowledge. Constructing crop portraits can provide a comprehensive description of crops and make up for these shortcomings. This research used agricultural question-and-answer data and popular science data obtained by text crawling as the original data, selected labels to establish a crop portrait that including three categories (crops, pesticides, and diseases and pests), and used the graph database (Neo4j) to store and display these portrait data. Information mining found that the crop portrait revealed the occurrence trend of diseases and pests, exhibited a nonintrinsic connection between different diseases and pests, and provided a variety of pesticides to choose from for control of diseases and pests. The results showed that constructing crop portraits is beneficial to agricultural analysis, and has practical application values and theoretical research prospects in the field of big data analytics.

Multi-Dimensional Event Data in Graph Databases

Process event data is usually stored either in a sequential process event log or in a relational database. While the sequential, single-dimensional nature of event logs aids querying for (sub)sequences of events based on temporal relations such as “directly/eventually-follows,” it does not support querying multi-dimensional event data of multiple related entities. Relational databases allow storing multi-dimensional event data, but existing query languages do not support querying for sequences or paths of events in terms of temporal relations. In this paper, we propose a general data model for multi-dimensional event data based on labeled property graphs that allows storing structural and temporal relations in a single, integrated graph-based data structure in a systematic way. We provide semantics for all concepts of our data model, and generic queries for modeling event data over multiple entities that interact synchronously and asynchronously. The queries allow for efficiently converting large real-life event data sets into our data model, and we provide 5 converted data sets for further research. We show that typical and advanced queries for retrieving and aggregating such multi-dimensional event data can be formulated and executed efficiently in the existing query language Cypher, giving rise to several new research questions. Specifically, aggregation queries on our data model enable process mining over multiple inter-related entities using off-the-shelf technology.

Distributed temporal graph analytics with GRADOOP

Temporal property graphs are graphs whose structure and properties change over time. Temporal graph datasets tend to be large due to stored historical information, asking for scalable analysis capabilities. We give a complete overview of Gradoop, a graph dataflow system for scalable, distributed analytics of temporal property graphs which has been continuously developed since 2005. Its graph model TPGM allows bitemporal modeling not only of vertices and edges but also of graph collections. A declarative analytical language called GrALa allows analysts to flexibly define analytical graph workflows by composing different operators that support temporal graph analysis. Built on a distributed dataflow system, large temporal graphs can be processed on a shared-nothing cluster. We present the system architecture of Gradoop, its data model TPGM with composable temporal graph operators, like snapshot, difference, pattern matching, graph grouping and several implementation details. We evaluate the performance and scalability of selected operators and a composed workflow for synthetic and real-world temporal graphs with up to 283 M vertices and 1.8 B edges, and a graph lifetime of about 8 years with up to 20 M new edges per year. We also reflect on lessons learned from the Gradoop effort.

A model and query language for temporal graph databases

Graph databases are becoming increasingly popular for modeling different kinds of networks for data analysis. They are built over the property graph data model, where nodes and edges are annotated with property-value pairs. Most existing work in the field is based on graphs were the temporal dimension is not considered. However, time is present in most real-world problems. Many different kinds of changes may occur in a graph as the world it represents evolves across time. For instance, edges, nodes, and properties can be added and/or deleted, and property values can be updated. This paper addresses the problem of modeling, storing, and querying temporal property graphs, allowing keeping the history of a graph database. This paper introduces a temporal graph data model, where nodes and relationships contain attributes (properties) timestamped with a validity interval. Graphs in this model can be heterogeneous, that is, relationships may be of different kinds. Associated with the model, a high-level graph query language, denoted T-GQL, is presented, together with a collection of algorithms for computing different kinds of temporal paths in a graph, capturing different temporal path semantics. T-GQL can express queries like “Give me the friends of the friends of Mary, who lived in Brussels at the same time than her, and also give me the periods when this happened”. As a proof-of-concept, a Neo4j-based implementation of the above is also presented, and a client-side interface allows submitting queries in T-GQL to a Neo4j server. Finally, experiments were carried out over synthetic and real-world data sets, with a twofold goal: on the one hand, to show the plausibility of the approach; on the other hand, to analyze the factors that affect performance, like the length of the paths mentioned in the query, and the size of the graph.