Domain Knowledge Model Research Articles

Unsupervised Anomaly Detection of Industrial Robots Using Sliding-Window Convolutional Variational Autoencoder

With growing dependence of industrial robots, a failure of an industrial robot may interrupt current operation or even overall manufacturing workflows in the entire production line, which can cause significant economic losses. Hence, it is very essential to maintain industrial robots to ensure high-level performance. It is widely desired to have a real-time technique to constantly monitor robots by collecting time series data from robots, which can automatically detect incipient failures before robots totally shut down. Model-based methods are typically used in anomaly detection for robots, yet explicit domain knowledge and accurate mathematical models are required. Data-driven techniques can overcome these limitations. However, a major difficulty for them is the lack of sufficient fault data of industrial robots. Besides, the used technique for anomaly detection of robots should be required to not only capture the temporal dependency in collected time series data, but also the inter-correlations between different metrics. In this paper, we introduce an unsupervised anomaly detection for industrial robots, sliding-window convolutional variational autoencoder (SWCVAE), which can realize real-time anomaly detection spatially and temporally by coping with multivariate time series data. This method has been verified by a KUKA KR6R 900SIXX industrial robot, and the results prove that the proposed model can successfully detect anomaly in the robot. Thus, this work presents a promising tool for condition-based maintenance of industrial robots.

Trepan Reloaded: A Knowledge-driven Approach to Explaining Artificial Neural Networks

Explainability in Artificial Intelligence has been revived as a topic of active research by the need of conveying safety and trust to users in the `how' and `why' of automated decision-making. Whilst a plethora of approaches have been developed for post-hoc explainability, only a few focus on how to use domain knowledge, and how this influences the understandability of global explanations from the users' perspective. In this paper, we show how ontologies help the understandability of global post-hoc explanations, presented in the form of symbolic models. In particular, we build on Trepan, an algorithm that explains artificial neural networks by means of decision trees, and we extend it to include ontologies modeling domain knowledge in the process of generating explanations. We present the results of a user study that measures the understandability of decision trees using a syntactic complexity measure, and through time and accuracy of responses as well as reported user confidence and understandability. The user study considers domains where explanations are critical, namely, in finance and medicine. The results show that decision trees generated with our algorithm, taking into account domain knowledge, are more understandable than those generated by standard Trepan without the use of ontologies.

Diversicon: Pluggable Lexical Domain Knowledge

Natural language understanding is a key task in a wide range of applications targeting data interoperability or analytics. For the analysis of domain-specific data, specialised knowledge resources (terminologies, grammars, word vector models, lexical databases) are necessary. The heterogeneity of such resources is, however, a major obstacle to their efficient use, especially in combination. This paper presents the open-source Diversicon Framework that helps application developers in finding, integrating, and accessing lexical domain knowledge, both symbolic and statistical, in a unified manner. The major components of the framework are: (1) an API and domain knowledge model that allow applications to retrieve domain knowledge through a common interface from a diversity of resource types, (2) implementations of the API for some of the most commonly used symbolic and statistical knowledge sources, (3) a domain-aware knowledge base that helps integrate static lexico-semantic resources, and (4) an online catalogue that either hosts or links to the existing resources from multiple domains. Support for Diversicon is already integrated into two of the most popular ontology matcher applications, a fact that we exploit to validate the framework and demonstrate its use on a example study that evaluates the effect of several common-sense and domain knowledge resources on a medical ontology matching task.

Integrated design environment for reusable modular assembly systems

PurposeThis paper aims to provide a service-based integrated prototype framework for the design of reusable modular assembly systems (RMAS) incorporating reusability of equipment into the process. It extends AutomationML (AML) developments for an engineering data exchange to integrate and standardize the data formats that support the design of RMAS.Design/methodology/approachThe approach provides a set of systematic procedures and support tools for the design of RMAS. This includes enhanced domain knowledge models that facilitate the interpretation and integration of information across the design phases.FindingsThe inclusion of reusability aspects in the design phase improves the sustainability of future assembly systems, by ensuring equipment use until its end-of-life. Moreover, the integrated support tools reduce the design time, while improving the quality/performance of the system design solution, as it enables the exploration of a larger solution space. This will result in a better response to dynamic and rapidly changing system requirements.Social implicationsThis work provides a sustainable approach for the design of modular assembly systems (MAS), which will ensure better resource utilization. Additionally, the standardization of the data and the support of low cost tools is expected to benefit industrial companies, particularly the small- and medium-sized enterprises.Originality/valueThis approach offers a service-based platform which uses production data to incorporate reusability aspects into the design process of modular assembly system. Moreover, it provides a framework for modular assembly system design by extending the current design processes and interactions between stakeholders. To support this, a standardized method for information representation and exchange across the several phases of the RMAS design activity is briefly illustrated with an industrial case study.

MACE: Mobile Artificial Conversational Entity for Adapting Domain Knowledge and Generating Personalized Advice

Mobile learning offers, with the help of handheld devices, a continuous access to the learning process. With the advent of mobile learning, educational systems are changing, offering the possibility of distance education without the restrictions of place and time. As such, new technological advancements are employed by mobile learning. This paper presents the design, development and evaluation of a novel artificial conversational entity, incorporated in a mobile learning system for personalized English language instruction. More specifically, it offers amelioration of the domain knowledge model by adapting it to the students’ needs and to the pace that they prefer to receive learning. Moreover, it creates personalized tutoring advice in order to support students in the educational process. Finally, it can assist the procedure of assessments since it automatically generates questions to assess the knowledge level of students. The evaluation of the mobile tutoring system presents promising results regarding the incorporation of this new technology in digital education with the aim of creating a student-centric learning experience.

ONCObc-ST: An Improved Clinical Reasoning Algorithm Based on Select and Test (ST) Algorithm for Diagnosing Breast Cancer

The need for an accurate reasoning algorithm is usually necessitated by the sensitivity of domain of (medicine as example) application of such algorithms. Most reasoning algorithms for medical diagnosis are either limited by their techniques or accuracy and efficiency. Even the Select and Test (ST) algorithm which is considered a more approximate reasoning algorithm is also limited by its approach of using bipartite graph in modeling domain knowledge and making inference through the use of orthogonal vector projection for estimating likelihood of diagnosis at the clinical decision stage (induction). While the bipartite graph knowledge base lacks n-ary use of predicate on concepts, orthogonal vector projection on the other hand has high computation for the inference process. The aim of this paper is to enhance ST algorithm for improved performance and accuracy. First, we propose the use of ontologies and semantic web based rule for knowledge representation so as to provide support for inference making. Furthermore, three major improvements were added to ST algorithm to aid the improvement of its approximation. Secondly, we designed an inference making procedure to enable interaction with the knowledge base mentioned earlier. Thirdly, we model Hill’s Criteria of Causation into clinical decision stage of ST to overcome the limitation of orthogonal vector projection. Lastly, the improved ST algorithm was largely represented and described using set notations (though implemented as linked-list and queues) and mathematical notations. The result of the improved ST algorithm revealed a sensitivity of 0.81 and 0.89 and specificity of 0.82 and 1.0 in the Wisconsin Breast Cancer Database (WBCD) and Wisconsin Diagnostic Breast Cancer (WDBC) datasets respectively. In addition, the accuracy obtained from the proposed algorithm was 86.0% and 88.72% for the Wisconsin Breast Cancer Database (WBCD) and Wisconsin Diagnostic Breast Cancer (WDBC) datasets respectively. This enhancement in accuracy was obtained at a slowdown time due to the reasoning process and ontology parsing task added to the enhanced system. However, there was an improvement in the accuracy and inference power of the resulting system.

A Method for Designing Domain-Specific Document Retrieval Systems using Semantic Indexing

Using domain knowledge and semantics to con-duct e‡ective document retrieval has attracted great attention from researchers in many di‡erent communities. Ultilizing that approach, we presents the method for designing domain-speciÿc document retrieval systems, which manages semantic information related to document content and supports se-mantic processing in search. ⁄e proposed method integrates components such as an ontology describing domain knowl-edge, a database of document repository, semantic represen-tations for documents; and advanced search techniques based on measuring semantic similarity. In this article, a model of domain knowledge for various information retrieval tasks, called ⁄e Classed Keyphrase based Ontology (CK-ONTO), will be presented in details. We also present graph-based models for representing documents together measures for evaluating the semantic relevance for usage in searching. ⁄e above methodology has been used in designing many real-world applications such as the Job-posting retrieval system. Evaluation with real-world inspired dataset, our methods showed noticeable improvements over traditional retrieval solutions.

DATA MINING OF COMMODITY PROPOSALS BASED ON CONTEXT RECOMMENDATIONS

Internet technologies are an integral part of the relationship that arises in modern society. The rapid introduction and convenience of electronic platforms triggered the projected growth in demand in the market for IT products for recommender systems.The article discusses various limitations of current recommender methods and discusses possible extensions that can improve the recommender capabilities and make them more valuable for a wide range of applications. These extensions include improving the perception of users and elements, including contextual information in the recommendatory process, supporting multi-criteria ratings and providing more flexible and at the same time less intrusive types of recommendations.When integrating the relevant information technology to develop a commodity proposals environment, it is therefore necessary to consider the personalization requirements of the proposal to ensure that the technology achieves its intended result. This study therefore sought to apply context aware technology and recommendation algorithms to develop a system realize personalized goals in a context aware manner and improve commodity proposals effectiveness.In order to offer context-aware and personalized information, intelligent processing techniques are necessary. Different initiatives considering many contexts have been proposed, but users preferences need to be learned to offer contextualized and personalized services, products or information. Therefore, this paper proposes an agent-based architecture for context-aware and personalized event recommendation based on ontology and the spreading algorithm. The use of ontology allows to define the domain knowledge model, while the spreading activation algorithm learns user patterns by discovering user interests.Also from the statistical observation, it is found that there exists a higher level agreement towards the system between the participants of both end users and experts.

Efficient management and storage of a multiversion OWL 2 DL domain ontology

AbstractOntology is the chief technology to model domain knowledge and fix its heterogeneity. Knowledge evolution is unavoidable in all fields, and ontology should reflect such an evolution while preserving its consistency. The access to an ontology evolution history is also a crucial need. Thereby, ontology versions should be stored efficiently. To address these requirements, some works focused only on managing ontology inconsistency. To this end, they adopted an a posteriori approach that checks inconsistency after its occurrence. Others have mainly focused on some ontology versioning aspects other than the storage and the consistency preservation issues. Specifically, the state of the art storage strategies perform well some typical queries at the cost of a high storage space. Strategies with low space overhead lose the main specificity of ontology: the knowledge semantics modelling. So, it is useful to address all these shortcomings. This paper presents an approach that allows generating consistent OWL 2 DL ontology versions in a priori way. It also proposes an efficient storage strategy that preserves the semantics aspect of ontology and performs well typical queries with a low storage space. A protégé plug‐in is developed to show the feasibility of the main ideas presented in this paper.

Ontology reuse for multiagent system development through pattern classification

SummaryOntologies play a crucial role in multiagent systems (MASs) development, especially for domain knowledge modeling, interaction specifications, and behavioral aspect representation. Domain‐specific ontologies can be developed in an ad hoc or systematic manner through the incorporation of ontology development steps on the basis of agent‐oriented methodologies. Developing such ontologies, however, is challenging because of the extensive amounts of knowledge and experience required. Moreover, since many ontologies cater for very specific domains, the question arises of whether some can be reused for faster systems development. This paper attempts to answer this question by proposing an ontology pattern classification scheme to allow the reuse of existing ontology knowledge for MAS development. Specifically, ontology patterns relevant to the design problem at hand are identified through the pattern classification scheme. These patterns are then reused and shared among agent software communities during the system development phase. The effectiveness of the proposed approach is validated using a restaurant‐finder MAS case study. Our findings suggest that utilization of the classified ontology patterns reduces development time and complexity when dealing with domain‐specific applications. The scheme also seems useful for software practitioners, where searching and reusing the patterns can easily be done during the analysis, design, and implementation of MAS development.

Formal Modelling of Ontologies : An Event-B based Approach Using the Rodin Platform

This paper reports on the results of the French ANR IMPEX research project dealing with making explicit domain knowledge in design models. Ontologies are formalised as theories with sets, axioms, theorems and reasoning rules. They are integrated to design models through an annotation mechanism. Event-B has been chosen as the ground formal modelling technique for all our developments. In this paper, we particularly describe how ontologies are formalised as Event-B theories.

Attack scenario reconstruction using intrusion semantics

Security information and event management (SIEM) systems receive a large number of alerts from different intrusion detection systems. They are expected, from these alerts, to make reliable and timely decisions regarding the types of ongoing attack scenarios and their priorities. However, the lack of an agreed-upon vocabulary for the representation of the domain knowledge makes it difficult for state-of-the-art SIEM systems to effectively manage these complex decisions. To overcome this problem, an ontology-based expert system approach can provide domain knowledge modeling as a foundation for disambiguation of meaning and automatic reasoning regarding ongoing attack scenarios. The proposed approach reconstructs attack scenarios by reasoning based on the evidences in the alert stream. The main idea of the proposed approach is to identify the causal relation between alerts using their similarity. This approach assumes that the similarity between two successive steps in an attack scenario is greater than that of two non-successive steps. Moreover, the similarity between the steps of the same attack scenario is greater than that between the steps of two different attack scenarios. The benefit of the proposed approach includes the fast and incremental reconstruction of known and unknown attack scenarios without expert intervention, which is an enormous step forward in developing expert and intelligent systems for cyber security. We evaluated the proposed technique by performing experiments on two known datasets: DARPA 2000 and MACCDC 2012. The results prove the advantages of the proposed approach with regard to completeness and soundness criteria.

Example-driven modeling: on effects of using examples on structural model comprehension, what makes them useful, and how to create them

We present a controlled experiment for the empirical evaluation of example-driven modeling (EDM), an approach that systematically uses examples for model comprehension and domain knowledge transfer. We conducted the experiment with 26 graduate (Masters and Ph.D. level) and undergraduate (Bachelor level) students from electrical and computer engineering, computer science, and software engineering programs at the University of Waterloo. The experiment involves a domain model, with UML class diagrams representing the domain abstractions and UML object diagrams representing examples of using these abstractions. The goal is to provide empirical evidence of the effects of suitable examples on model comprehension, compared to having model abstractions only, by having the participants perform model comprehension tasks. Our results show that EDM is superior to having model abstractions only, with an improvement of 39% for diagram completeness, 33% for questions completeness, 71% for efficiency, and a reduction in the number of mistakes by 80%. We provide qualitative results showing that participants receiving model abstractions augmented with examples experienced lower perceived difficulty in performing the comprehension tasks, higher perceived confidence in their tasks’ solutions, and asked 90% fewer clarifying domain questions. We also present participants’ feedback regarding the usefulness of the provided examples, their number and types, as well as the use of partial examples. We present a taxonomy of the different types of examples, explain their significance, and propose guidelines for manual and automatic creation of useful examples.

Human resource optimisation through semantically enriched data

The industrial domain is experiencing a so-called fourth industrial revolution in which the evergrowing complexity of manufacturing information, the increasing amount of knowledge and the use of web-oriented techniques, represent three crucial factors that are accelerating the growth of complexity of industrial systems. On the other hand, continuous-evolving requirements in industrial environments, due to technology outbreaks and a new global marketplace, have led to an on-going evolution of human resource management through the creation and adoption of alternative business models. In the past decade, semantic models such as ontologies have been proven to be effective for many knowledge-intensive applications, since they provide formal models of domain knowledge that can be exploited in different ways. For all these reasons, an innovative human resource optimisation (HRO) engine is introduced, which employs semantically enhanced information and conditional random field (CRFs) probabilistic models with knowledge derived from industrial shop floor level, and proposes the right person for the right job in real-time shop floor operations towards optimising decisions on how to implement and schedule either repeatedly or non-occurring tasks. Industrial information data flow and semantic enrichment were ensured through the combined use of a common interface data exchange model (CIDEM) and ontologies, after which a feasibility study at a chemical plant presented interesting preliminary results.

A Java software agent framework for hard real-time manufacturing control

Abstract Multiagent systems (MAS) have widely been recognized as a suitable software engineering approach to design distributed, flexible, and robust control-systems, as are needed to cope with current and future challenges in manufacturing. Yet, applying MAS for real-time control has been subject to several concessions so far. This paper presents a real-time Java multiagent platform, which allows both agent execution and distributed interaction under hard real-time conditions. The paper covers the architecture and implementation of such a platform, the integration of a domain knowledge model into the communication flow, and finally an analytical response time model including interactions, to actually proove the real-time capability of a distributed MAS.

Combination of Data-driven Feature Selection Methods with Domain Knowledge for Diagnosis of Railway Vehicles

Railway vehicles are generally maintained preventively within certain time periods. Condition based predictive maintenance strategies have a great economic potential so that modern trains are equipped with many sensors in order to perform diagnostics and prognostics of components.Methods for fault detection need appropriate feature subsets in order to achieve small in-sample and out-sample errors. In our case the typical feature selection approach using pure data-driven methods is difficult, as the number of possible feature sets is very large. On the other hand there exists rich domain knowledge and detailed physical models of the mechanical system. The aim is to combine this knowledge with the often used mathematical methods for feature selection for improving classification of cases when a faulty damper is present. Based on the dynamic equations of motion, this paper presents heuristic feature selection via the analysis of transfer functions. We describe several wellknown methods of automated feature selection and a workflow which combines domain knowledge with automated methods. Results show that it is difficult to definefeatures based only on domain-knowledge, but in combination with data-driven techniques good classification performance can be achieved.

Development of a Functionality Matrix for a Cognitive Assistant on Long Distance Space Missions

This paper discusses the development and application of a Functionality Matrix (FM) technique to design requirements for a Cognitive Assistant (CA) to aid future long distance space missions. Future far-Earth space missions pose challenges due to communication delays between Mission Control and space crew. As communication delays increase, Mission Control is less able to support space crew in off-nominal situations. The FM technique was used to extrapolate the joint decision-making functions performed in current near-Earth operations to functions performed in future, far-Earth space operations, allocated between Mission Control, space crew, and potential automation. Data gathered from document analysis and interviews with astronauts were consolidated to generate the domain knowledge models of the collaborative decision-making process for current space operations. Each function in the decision-making process was examined to extrapolate the current collaborative work of Mission Control and space crew to future space operations under long time delays, when the role of Mission Control must change. The domain knowledge models were also used to identify the functions performed to resolve off-nominal situations in current operations that are most affected by long communication delays. The identified functions constituted the input for the FM. The responsibilities of Mission Control, space crew and CA were categorized in the FM as current and/or future space operations. The FM describes the shared responsibilities between space crew and Mission Control with and without CA support in space operations. The FM was used to describe functional design requirements, and will lead to the development of function allocation rules.

Automatic Ontology Construction Using Text Corpora and Ontology Design Patterns (ODPs) in Alzheimer’s Disease

An ontology is defined as an explicit specification of a conceptualization, which is an important tool for modeling, sharing and reuse of domain knowledge. However, ontology construction by hand is a complex and a time consuming task. This research presents a fully automatic method to build bilingual domain ontology from text corpora and ontology design patterns (ODPs) in Alzheimer’s disease. This method combines two approaches: ontology learning from texts and matching with ODPs. It consists of six steps: (i) Term & relation extraction (ii) Matching with Alzheimer glossary (iii) Matching with ontology design patterns (iv) Score computation similarity term & relation with ODPs (v) Ontology building (vi) Ontology evaluation. The result of ontology composed of 381 terms and 184 relations with 200 new terms and 42 new relations were added. Fully automatic ontology construction has higher complexity, shorter time and reduces role of the expert knowledge to evaluate ontology than manual ontology construction. This proposed method is sufficiently flexible to be applied to other domains.

Modeling travel mode choice behavior with bounded rationality using Markov Logic Networks

Disaggregate choice models have been widely studied to quantify the influence of the characteristics of travelers as well as the attributes of alternatives and choices in their travel modes. However, due to their model specifications and primary assumptions on unobserved disturbances, their modeling capability is constrained. In this study, a Markov Logic Network (MLN)-based approach is developed to combine bounded rationality principles with travelers’ behavior in travel mode choices. This approach is established based on logical domain knowledge and probabilistic models. MLN can extract logical domain knowledge and represent the impacts of significant attributes using independent logical formulas that are weighted correspondingly by their relative relationships. Travel-mode choice is determined based on travelers’ personal preferences and logical domain knowledge. The numerical examples and parameter sensitivity analyses indicate this approach performs reasonably well. The research findings are helpful to better understand travel mode-choice model specifications and travel behavior interpretations.

Understanding the complexity of axiom pinpointing in lightweight description logics

Lightweight description logics are knowledge representation formalisms characterised by the low complexity of their standard reasoning tasks. They have been successfully employed for constructing large ontologies that model domain knowledge in several different practical applications. In order to maintain these ontologies, it is often necessary to detect the axioms that cause a given consequence. This task is commonly known as axiom pinpointing.In this paper, we provide a thorough analysis of the complexity of several decision, counting, and enumeration problems associated to axiom pinpointing in lightweight description logics.