KEOD 2021 Abstracts


Full Papers
Paper Nr: 2
Title:

Health Ontology for Minority Equity (HOME)

Authors:

Navya M. Kollapally, Yan Chen and James Geller

Abstract: Healthcare inequity, as defined by the World Health Organization (WHO), is a systemic difference in healthcare services received by different population groups, based on race, ethnicity, gender, sexual orientation, etc. The Covid-19 pandemic has heightened the awareness of differences in care received by racial and ethnic minorities in the US. We have investigated the physical, psychological, and emotional harm that people of colour were exposed to during this time. It is necessary to record data about unequal treatment to identify and eradicate existing institutional racism in healthcare. Electronic Health Records (EHRs) rely to a high degree on “coded” terms from terminologies and ontologies. Such a biomedical ontology can be used for standardization, integration and sharing of data, knowledge reuse, decision support, etc. No ontology for racial differences exists in US healthcare. This motivation leads us to the development of such an ontology to record the physical, emotional, and psychological effects resulting from differences in treatment that citizens receive, based on their identity. Differences exist not only inside of healthcare organizations, but also occur even before entering them. We present the first version of such a Health Ontology for Minority Equity (HOME) along with ontology evaluation methods that we applied.
Download

Paper Nr: 8
Title:

An Evaluation of Hubness Reduction Methods for Entity Alignment with Knowledge Graph Embeddings

Authors:

Daniel Obraczka and Erhard Rahm

Abstract: The heterogeneity of Knowledge Graphs is problematic for conventional data integration frameworks. A possible solution to this issue is using Knowledge Graph Embeddings (KGEs) to encode entities into a lower-dimensional embedding space. However, recent findings suggest that KGEs suffer from the so-called hubness phenomenon. A dataset that suffers from hubness has a few popular entities that are nearest neighbors of a highly disproportionate amount of other entities. Because the calculation of nearest neighbors is an integral part of entity alignment with KGEs, hubness reduces the accuracy of the matching result. We therefore investigate a variety of hubness reduction techniques and utilize approximate nearest neighbor (ANN) approaches to offset the increase in time complexity stemming from the hubness reduction. Our results suggest, that hubness reduction in combination with ANN techniques improves the quality of nearest neighbor results significantly compared to using no hubness reduction and exact nearest neighbor approaches. Furthermore, this advantage comes without losing the speed advantage of ANNs on large datasets.
Download

Paper Nr: 10
Title:

Matching Entities from Multiple Sources with Hierarchical Agglomerative Clustering

Authors:

Alieh Saeedi, Lucie David and Erhard Rahm

Abstract: We propose extensions to Hierarchical Agglomerative Clustering (HAC) to match and cluster entities from multiple sources that can be either duplicate-free or dirty. The proposed scheme is comparatively evaluated against standard HAC as well as other entity clustering approaches concerning efficiency and efficacy criteria. All proposed algorithms can be run in parallel on a distributed cluster to improve scalability to large data volumes. The evaluation with diverse datasets shows that the new approach can utilize duplicate-free sources and achieves better match quality than previous methods.
Download

Paper Nr: 15
Title:

A Scalable Approach for Distributed Reasoning over Large-scale OWL Datasets

Authors:

Heba Mohamed, Said Fathalla, Jens Lehmann and Hajira Jabeen

Abstract: With the tremendous increase in the volume of semantic data on the Web, reasoning over such an amount of data has become a challenging task. On the other hand, the traditional centralized approaches are no longer feasible for large-scale data due to the limitations of software and hardware resources. Therefore, horizontal scalability is desirable. We develop a scalable distributed approach for RDFS and OWL Horst Reasoning over large-scale OWL datasets. The eminent feature of our approach is that it combines an optimized execution strategy, pre-shuffling method, and duplication elimination strategy, thus achieving an efficient distributed reasoning mechanism. We implemented our approach as open-source in Apache Spark using Resilient Distributed Datasets (RDD) as a parallel programming model. As a use case, our approach is used by the SANSA framework for large-scale semantic reasoning over OWL datasets. The evaluation results have shown the strength of the proposed approach for both data and node scalability.
Download

Paper Nr: 19
Title:

Hierarchical Control of Swarms during Evacuation

Authors:

Kristýna Janovská and Pavel Surynek

Abstract: The problem of evacuation is addressed from the perspective of agent-based modeling (ABM) in this paper. We study evacuation as a problem of navigation of multiple agents in a known environment. The environment is divided into a danger and a safe zone while the task of agents is to move from the danger zone to the safe one. Unlike previous approaches that model the environment as a discrete graph with agents placed in its vertices our approach adopts various continuous aspects such as grid-based embedding of the environment into 2D space continuous line of sight of an agent. In addition to this, we adopt hierarchical structure of multi-agent system in which so called leading agents are more informed and are capable of performing multi-agent pathfinding (MAPF) via centralized algorithms like conflict-based search (CBS) while so called follower agents are modeled using simple local rules. Our experimental evaluation indicates that suggested modeling approach can serve as a tool for studying the progress and the efficiency of the evacuation process.
Download

Paper Nr: 22
Title:

COMET: An Ontology Extraction Tool based on a Hybrid Modularization Approach

Authors:

Bernabé Batchakui, Emile Tawamba and Roger Nkambou

Abstract: The design of ontologies is a non-trivial task that can simply be reduced to the reuse of one or more existing ontologies. However, since an expert in knowledge engineering would only need a part of the ontology to perform a specific task, obtaining this partition will require the modularization of ontologies. This article proposes a tool named COMET, based on hybrid modularization, composed of existing structural and semantic modularization techniques, that, from an ontology and a list of input terms, generates, according to an integrated segmentation algorithm, a module which in fact is a segment consisting only of concepts deemed relevant. The segmentation algorithm is based on two parameters which are hierarchical deep and semantic threshold.
Download

Paper Nr: 23
Title:

Modelling and Detection of Driver’s Fatigue using Ontology

Authors:

Alexandre Lambert, Manolo Dulva Hina, Celine Barth, Assia Soukane and Amar Ramdane-Cherif

Abstract: Road accidents have become the eight leading cause of death all over the world. Lots of these accidents are due to a driver’s inattention or lack of focus, due to fatigue. Various factors cause driver’s fatigue. This paper considers all the measureable data that manifest driver’s fatigue, namely those manifested in the vehicle measureable data while driving as well as the driver’s physical and physiological data. Each of the three main factors are further subdivided into smaller details. For example, the vehicle’s data is composed of the values obtained from the steering wheel’s angle, yaw angle, the position on the lane, and the speed and acceleration of the vehicle while moving. Ontological knowledge and rules for driver fatigue detection are to be integrated into an intelligent system so that on the first sign of dangerous level of fatigue is detected, a warning notification is sent to the driver. This work is intended to contribute to safe road driving.

Paper Nr: 28
Title:

An Ontology based Task Oriented Dialogue

Authors:

João Q. Silva, Dora Melo, Irene P. Rodrigues, João C. Seco, Carla Ferreira and Joana Parreira

Abstract: An ontology based task oriented dialogue as an interface to an user developing applications using natural language instructions is presented. The dialogue system represents the domain knowledge in an OWL2 ontology that is consulted and updated in the interpretation of the user utterances. The utterances are processed using an Universal Dependencies parser whose output is then interpreted to obtain a partial semantic representation. The pragmatic interpretation computes a set of possible interpretations by matching the partial representation with the ontology information, classes, properties, instances and data properties values, such as names. The dialogue manager is able to use soft constraints to choose the set of best interpretations. A set of preliminary experimental cases with promising results is also presented.
Download

Short Papers
Paper Nr: 3
Title:

Applicability of a Foundational Ontology to Semantically Enrich the Core and Domain Ontologies

Authors:

Luis Olsina

Abstract: This paper analyses the key terms, relationships and axioms of ThingFO (Thing Foundational Ontology), which is an ontology devoted for particular and universal things and assertions. It is placed at the foundational level in the context of a five-tier ontological architecture. This architecture groups together foundational, core, top-domain, low-domain, and instance levels, making ThingFO the single ontology at the top level. Thus, the ontologies at lower levels reuse and specialize, for example, its terms and relationships. To illustrate the applicability of ThingFO, this work also discusses enriched terms and specialized relationships for a core ontology, particularly for situation, where its concepts are themselves cross-cutting concerns for different domain terminologies. In addition, verification and validation issues are addressed as well.
Download

Paper Nr: 4
Title:

Support in Policymaking: A Systematic Exploration of the Policymaking Process

Authors:

Daniel Guzman Vargas and Sidharta Gautama

Abstract: Nearly all the public policy issues focus on complex social problems (sometimes referred to as ‘wicked’ problems). Failing to address such complexity may result in a weak formulation of the problem at hand and consequently to policy failure. A decision support system (DSS) appropriate for handling ẁicked' problems in policymaking should help decision-makers cope with the problem's complexity, facilitate the assessing of multiple alternatives, and favour a discussion towards a common agenda. Making use of the above requirements, we present in this paper a methodology for a DSS that feeds from a frame representation of both expert knowledge and policy-related evidence to support decision-makers in the policymaking process. The application of the methodology in a specific use case suggests the methodology could be applied in a DSS for the identification of patterns and trends in policy-relevant data, the identification of possible policy configurations, and the drafting of alternative scenarios based on the possible configurations.
Download

Paper Nr: 6
Title:

An Object Oriented Approach for Ontology Modeling and Reasoning

Authors:

Ouassila L. Narsis and Christophe Nicolle

Abstract: In Industry, IT personnel commonly manipulate the object paradigm for software engineering. Industry 4.0 with its profusion of data requires the implementation of a knowledge engineering approach, breaking with data processing habits. The way from an object practice to an ontological vision is a gap for most of the staff, and transition from a business expert model to software implementation is the source of many errors and delays. Moreover, in ontology engineering, namely that domain experts have significant difficulties learning ontology languages and correctly using them. To overcome this problem, this paper presents the first results of a modeling and reasoning approach based on the object paradigm by combining UML modeling and reasoning by constraint resolution with OCL. The ontology produced by domain experts can then be checked and improved by a UML/OCL approach which is easily understandable and often more familiar to engineers.
Download

Paper Nr: 7
Title:

Towards Automation of Regulatory Compliance Checking in the Product Design Phase

Authors:

Malte Ramonat, Andreas W. Müller and Alexander Fay

Abstract: The process of checking if a designed product is compliant with standards is time-consuming and error-prone. This paper presents an approach for the automation of compliance checking using tables and formulae of standards as information sources. An ontology is created to enable comparisons between parameter values specified in standards in the form of a PDF document and parameter values of a designed product saved in a 3D PDF document. The extraction of regulatory information from PDF documents is discussed and software tools for information extraction are compared.
Download

Paper Nr: 9
Title:

Stacking BERT based Models for Arabic Sentiment Analysis

Authors:

Hasna Chouikhi, Hamza Chniter and Fethi Jarray

Abstract: Recently, transformer-based models showed great success in sentiment analysis and were considered as the state-of-the-art model for various languages. However, the accuracy of Arabic sentiment analysis still needs improvements. In this work, we proposed a stacking architecture of Arabic sentiment analysis by combing different BERT models. We also create a large-scale dataset of Arabic sentiment analysis by merging small publicly available datasets. The experimental study proves the efficiency of the proposed approach in terms of classification accuracy compared to single model architecture.
Download

Paper Nr: 11
Title:

Ontology-based Framework for Integration of Time Series Data: Application in Predictive Analytics on Data Center Monitoring Metrics

Authors:

Lauri Tuovinen and Jaakko Suutala

Abstract: Monitoring a large and complex system such as a data center generates many time series of metric data, which are often stored using a database system specifically designed for managing time series data. Different, possibly distributed, databases may be used to collect data representing different aspects of the system, which complicates matters when, for example, developing data analytics applications that require integrating data from two or more of these. From the developer’s point of view, it would be highly convenient if all of the required data were available in a single database, but it may well be that the different databases do not even implement the same query language. To address this problem, we propose using an ontology to capture the semantic similarities among different time series database systems and to hide their syntactic differences. Alongside the ontology, we have developed a Python software framework that enables the developer to build and execute queries using classes and properties defined by the ontology. The ontology thus effectively specifies a semantic query language that can be used to retrieve data from any of the supported database systems, and the Python framework can be set up to treat the different databases as a single data store that can be queried using this semantic language. This is demonstrated by presenting an application involving predictive analytics on resource usage and electricity consumption metrics gathered from a Kubernetes cluster, stored in Prometheus and KairosDB databases, but the framework can be extended in various ways and adapted to different use cases, enabling machine learning research using distributed heterogeneous data sources.
Download

Paper Nr: 12
Title:

Building Operating Systems: A Cloud-based Architecture for Enabling Knowledge Representation and Improving the Adaptability in Smart Buildings

Authors:

Adrian Taboada-Orozco, Kokou Yetongnon and Christophe Nicolle

Abstract: Hardware and software’s constant evolution opens new horizons for developing systems focused on buildings. Thanks to this evolution, the smartization of them is possible. The target of building automation systems is always to reduce human intervention, thus avoiding errors and maximizing the building resources. However, these systems still require an operator that knows the building’s topology, the elements inside, the dwellers’ preferences, and how to operate the automation system. The operator uses this knowledge to adapt the automation system by configuring parameters and thresholds to deal with the most common issues in buildings. This paper introduces a new approach named WITTYM-BOS (W-BOS) that consists of a unique Building Operating System (BOS) architecture, doted with a knowledge representation comparable with the operators’ knowledge. In this way, we transfer the abilities of an operator to improve the adaptability of a building automation system like BOS. To construct the knowledge, we combine static and dynamic information such as Industrial Foundation Classes (IFC) and data coming from the Internet of Things (IoT). We construct a preliminary prototype to illustrate our concept and validate use cases. Our work opens new horizons to innovative applications that profit from the easy understandability of our approach.
Download

Paper Nr: 18
Title:

OntoAqua: Ontology-based Modelling of Context in Water Safety and Security

Authors:

Alexandros-Michail Koufakis, Savvas Tzanakis, Anastasia Moumtzidou, Georgios Meditskos, Anastasios Karakostas, Stefanos Vrochidis and Ioannis Kompatsiaris

Abstract: Water distribution systems are comprised of a variety of different components that must be monitored in order to combat crises as effectively as possible. In particular, the subsystems that monitor the different components are varied and diverse, and as a result, their produced data are heterogeneous and occupy different modalities. This paper describes the OntoAqua ontology that aims to semantically represent knowledge and data sources in the event of a water-related crisis, including preparatory and follow-up measures. Towards the creation of an ontology that is semantically sound and adopts international standards, existing ontologies and resources were reused. More specifically, the specification and the semantics of the ontology are inspired mainly from the ISO 15975 - Security of drinking water supply. The modelling of sensor data was implemented by reusing the SAREF ontology and its extension for the water domain. For crowdsourcing and social media, the ontology imports classes and properties from the SIOC ontology.
Download

Paper Nr: 29
Title:

Neuro-symbolic XAI for Computational Drug Repurposing

Authors:

Martin Drancé, Marina Boudin, Fleur Mougin and Gayo Diallo

Abstract: Today in the health domain, the challenge is to build a more transparent artificial intelligence, less affected by the opacity intrinsic to the mathematical concepts it uses. Among the fields which use AI techniques, is drug development, and more specifically drug repurposing. DR involves finding a new indication for an existing drug. The hypotheses generated by DR techniques must be validated. Therefore, the mechanism of generation must be understood. In this paper, we describe the use of a state-of-the-art neuro-symbolic algorithm in order to explain the process of link prediction in a knowledge graph-based computational drug repurposing. Link prediction consists of generating hypotheses about the relationships between a known molecule and a given target. More specifically, the implemented approach allows to understand how the organization of data in a knowledge graph changes the quality of predictions.
Download

Paper Nr: 32
Title:

An Ontology-based Approach to Social Networks Mining

Authors:

Viacheslav Lanin, Lyudmila Lyadova, Elena Zamyatina and Nikita Vostroknutov

Abstract: The article presents an approach to the analysis of processes in social networks based on using multifaceted ontologies. An overview of existing tools for analyzing social networks is provided and the results of studying social networks are presented. A multifaceted ontology describing social networks has been developed based on the research findings. The main result for this study is the ontology of events, which can be used to pre- process data, extracted from social networks, to generate event logs in a form suitable for export to Process Mining tools to analyze networks (identification of user behavior patterns, analysis of information distribution, etc.). Examples illustrating the proposed approach are given.
Download

Paper Nr: 13
Title:

Representing BORM Process Models using OWL and RDF

Authors:

Marek Suchánek and Robert Pergl

Abstract: Business Object Relationship Modeling (BORM) is a business process analysis method based on communicating finite-state machines and Petri nets. However, because of its tooling support and non-interoperability of formats, it is rather niche. This work proposes a way of representing the knowledge from BORM process models in RDF by creating a BORM ontology. It benefits from previous work done on different conceptual and process modelling languages and their transformations to OWL. The resulting RDF representation brings increased interoperability and enhanced analysis possibilities, e.g., using SPARQL or RDF visualization tools. A part of this work is also a BORM-to-RDF export feature for the OpenPonk modelling platform. The resulting BORM ontology is ready for use in practice and further work.
Download

Paper Nr: 16
Title:

Specialized Neural Network Pruning for Boolean Abstractions

Authors:

Jarren Briscoe, Brian Rague, Kyle Feuz and Robert Ball

Abstract: The inherent intricate topology of a neural network (NN) decreases our understanding of its function and purpose. Neural network abstraction and analysis techniques are designed to increase the comprehensibility of these computing structures. To achieve a more concise and interpretable representation of a NN as a Boolean graph (BG), we introduce the Neural Constantness Heuristic (NCH), Neural Constant Propagation (NCP), shared logic, the Neural Real-Valued Constantness Heuristic (NRVCH), and negligible neural nodes. These techniques reduce a neural layer’s input space and the number of nodes for a problem in NP (reducing its complexity). Additionally, we contrast two parsing methods that translate NNs to BGs: reverse traversal (N ) and forward traversal (F ). For most use cases, the combination of NRVCH, NCP, and N is the best choice.
Download

Paper Nr: 17
Title:

Questions and Answers in Parliamentary Discussions: Form and Functions

Authors:

Mare Koit

Abstract: The study is aimed to develop the Estonian parliamentary corpus. The existing morphologically analyzed corpus includes verbatim records of sessions held in the Parliament of Estonia in 1995-2001. An important task of the Parliament is the passing of acts and resolutions. Every reading of a bill starts with a speech of a minister and/or of a member of the responsible leading committee. Then members of the Parliament can ask questions which will be answered by the presenter. The paper concentrates on the questions and answers that have been annotated in the corpus according to a custom-made dialogue act annotation scheme as well as the ISO standard. For comparison, questions and answers when reading a bill in the UK Parliament are considered. Different forms of questions and answers with different functions are prevailing in both parliaments. The main function of questions in the Parliament of Estonia is to get information. On the contrary, in the UK Parliament the questions mainly are used to present arguments for or against the bill. The main function of answers is to provide information in the Parliament of Estonia but agreement or disagreement with arguments in the UK Parliament. Our further aim is the automatic analysis of Estonian political texts and comparison with political discourse in other parliaments.
Download

Paper Nr: 20
Title:

Semi-Structured Schema for a Big Data (S-SSBD)

Authors:

Shady Hamouda, Raed Sughayyar and Omar Elejla

Abstract: Big data has become a crucial issue and has emerged as one of the most important technologies in the modern world. One of the concerns that need to be addressed regarding big data is the lack of a method that can handle a semi-structured data model with a flexible schema. None of the existing semi-structured models can handle a large volume of data with a flexible schema. Therefore, these requirements give significance to designing and develop a schema for semi-structured data. In addition, these issues and challenges have to be addressed by researchers when designing a method or algorithm to retrieve information from a large amount of data. This study aims to design a semi-structured data schema for big data applications. This study assessed the flexible schema feature and data based on the semi-structure features. The result showed that the proposed schema can cover and handle any change in the schema.
Download

Paper Nr: 26
Title:

Smart Lifts: An Ontological Perspective

Authors:

D. Slee, S. Cain, P. Vichare and J. I. Olszewska

Abstract: Nowadays, there is a growth of smart factories and Industry 4.0 technologies, involving Artificial Intelligence (AI) systems. These ones require interoperable solutions. In particular, ontologies have been widely used for capturing, sharing, and representing knowledge in an interoperable way, that both humans and machines can understand. Indeed, ontologies allow humans to communicate with machines in a semantic way, while machines are able to make automated reasoning about the concepts and relationships which are encoded in the ontology. For this purpose, this paper proposes the first-ever domain ontology for smart lifts. Its domain covers smart lift design, operation, and maintenance, while its scope is to aid in automating such lift services. This smart lift ontology (SLO), which contains 144 classes and 749 axioms, has been successfully developed in collaboration with the elevator industry.
Download

Paper Nr: 31
Title:

Ontology for the Semantic Enhancement, Database Definition and Management and Revision Control

Authors:

Edward S. Blurock

Abstract: This paper describes the use of ontologies interacting with a noSQL database (Google Cloud Firestore) in multiple capacities in the database system CHEMCONNECT. The motivation is to implement the ‘Data on the Web Best Practices” as recommended by the W3C (https://www.w3.org/TR/2017/REC-dwbp-20170131/, 2017) in an application within the physical chemistry and instrumentation. First, the ontology provides semantic enhancement to each database object through meta-data, standard vocabularies and data object relationships. There is a one-to-one correspondence between the database objects and the ontology objects. Another use of the ontology is to provide a data-driven model for the creation, provenance and versioning of database objects. One aspect of this is the use of domain specific templates to guide the construction of the database objects. The definition of each database object is in a hierarchy of catalog objects, record objects and components (using the DCAT ontology model). Within each of these object definitions is a link describing how a create a set of automatically generated RDF objects within the CHEMCONNECT database. The RDFs facilitate searching the database. To facilitate versioning, data source tracking and data quality control, operations on the database are organized as transactions. In CHEMCONNECT a transaction has a one to one correspondence with the underlying JAVA operation in the implementation. Within the transaction definition, the set of prerequisites and the output of the operation is defined. The use of transactions helps organize and give semantic enhancement to the set of individual operations within the implementation. The work in this paper is on-going and as the first use-case is concentrating experimental and theoretical information in the chemical domain. The implementation is written in JAVA and is using Google Cloud firestore as the database.
Download