Domain-specific Modeling Languages Research Articles

ProcessBERT: A Pre-trained Language Model for Judging Equivalence of Variable Definitions in Process Models*

Digital twins are expected to play a pivotal role in digital transformation. Although process informatics has attracted much attention, physical models are essential to realizing the digital twins. However, building a physical model of an industrial process takes much toil. We aim to facilitate the physical model building by developing an automated physical model building AI, named AutoPMoB, which performs five tasks: 1) retrieving documents about a target process from literature databases, 2) converting the format of each document to HTML format, 3) extracting information required for building a physical model from the documents, such as variables, equations, and experimental data, 4) judging the equivalence of the information extracted from different documents, and 5) reorganizing the information to output a desired physical model. This study focuses on task 4, especially judging the equivalence of variable definitions, i.e., whether two noun phrases represent the same variable. We created a large-scale corpus consisting of papers on chemical engineering, and built ProcessBERT, which is a domain-specific language model pre-trained on the corpus. We proposed a method for judging the equivalence of variable definitions based on ProcessBERT. When judging the equivalence, our proposed method first uses ProcessBERT to obtain the embeddings of the variable definitions. Then, the method calculates the cosine similarity between the embeddings. The method judges that the two definitions are equivalent when the similarity is larger than a threshold. Our proposed method judged the equivalence with higher accuracy than the method based on original BERT and SciBERT.

Domain-Specific Language Model Pre-Training for Korean Tax Law Classification

Owing to their increasing amendments and complexity, most taxpayers do not have the required knowledge of tax laws, which results in issues in everyday life. To use tax counseling services through the internet, a person must first select a category of tax laws corresponding to their tax question. However, a layperson without prior knowledge of tax laws may not know which category to select in the first place. Therefore, a model capable of automatically classifying the categories of tax laws is needed. Recently, a model using BERT has been frequently used for text classification; however, it is generally used in open-domains, and often experiences a degraded performance due to domain-specific technical terms, such as tax laws. Furthermore, a significant amount of time is required to train the model, since BERT is a large-scale model. To address these issues, this study proposes Korean tax law-BERT (KTL-BERT) for the automatic classification of categories of tax questions. For the proposed KTL-BERT, a new pre-trained language model was constructed by performing learning from scratch, to which a static masking method was applied based on DistilRoBERTa. Subsequently, the pre-trained language model was fine-tuned to classify five categories of tax law. A total of 327,735 tax law questions were used to verify the performance of the proposed KTL-BERT. The F1-score of the proposed KTL-BERT was approximately 91.06%, which is higher than that of the benchmark models by approximately 1.07%-15.46%, and the training speed was approximately 0.89%-56.07% higher.

Design Methodology and Tools in Factory of the Future

Abstract This paper presents a design method and tool developed to support the skill forming activities in the DigiFoF network (https://www.digifof.eu/). The focus is on training of manufacturing system design skills both as HEI education and vocational training, but preliminary design of new manufacturing systems is also supported (e.g in the development of small business process scenarios). We proposed a model-based methodology for solving of the manufacturing system design problems The methodology and the supporting tool are centred around a less abstract Domain-Specific Modelling Language (DSML). The language is easy to learn due to its few components. A modelling and simulation environment named Digital Production Planner Tool (DPPT) was generated from the metamodel of the DSML. The degree of abstraction used by this tool corresponds well to the intended use in training and preliminary design. Our method incorporates by design the possibility to impose constraints at the modelling language level to limit the modelling space to feasible/possible solutions. The resulting tool enforces these constraints in the use and supports the development of feasible designs even by inexperienced designers. The access to the conceptual model allows the translation of the model to other modelling language like Petri net. This extends the support for the design methodology. The whitepaper presents a use case for the developed method and tool: the design of a chocolate manufacturing line.

A domain-specific language for modeling and analyzing solution spaces for technology roadmapping

The introduction of major innovations in industry requires a collaboration across the whole value chain. A common way to organize such a collaboration is the use of technology roadmaps, which act as an industry-wide long-term planning tool. Technology roadmaps are used to identify industry needs, estimate the availability of technological solutions, and identify the need for innovation in the future. Roadmaps are inherently both time-dependent and based on uncertain values, i.e., properties and structural components can change over time. Furthermore, roadmaps have to reason about alternative solutions as well as their key performance indicators. Current approaches for model-based engineering do not inherently support these aspects.We present a novel model-based approach treating those aspects as first-class citizens. To address the problem of missing support for time in the context of roadmap modeling, we introduce the concepts of a common global time, time-dependent properties, and time-dependent availability. This includes requirements, properties, and the structure of the model or its components as well. Furthermore, we support the specification and analysis of key performance indicators for alternative solutions. These concepts result in a continuous range of various valid models over time instead of a single valid model at a certain point of time. We present a graphical user interface to enable the user to efficiently create and analyze those models. We further show the semantics of the resulting model by a translation into a set of global constraints as well as how we solve the resulting constraint system. We report on the evaluation of these concepts and the Iris tool with domain experts from different companies in the automotive value chain based on the industrial case of a smart sensing electrical fuse.

Multi-level production process modeling language

The fourth industrial revolution introduces changes in traditional manufacturing systems and creates basis for a lot-size-one production. The complexity of production processes is significantly increased, alongside the need to enable efficient process simulation, execution, monitoring, real-time decision making and control. The main goal of our research is to define a methodological approach and a software solution in which the Model-Driven (MD) principles and Domain-Specific Modeling Languages (DSMLs) are used to create a framework for the formal description and automatic execution of production processes. In that way production process models are used as central artifacts to manage the production. In this paper, we analyze production process modeling domain and present a DSML which can be used to create production process models suitable for automatic generation of executable code. The generated code is used for automatic execution of production processes within a simulation or a shop floor. The language can be used to specify errors that may occur during the process execution and to specify error handling and corrective steps, too. The DSML is evaluated by different groups of users and the evaluation results are presented. Both the DSML and the accompanying modeling tool are still in the prototype phase, as they are created and evaluated in use cases covering just the assembly of goods. To enable wider application of the language and the tool, it is required to have additional use cases from different manufacturing domains.

Medical Named Entity Recognition from Un-labelled Medical Records based on Pre-trained Language Models and Domain Dictionary

Medical named entity recognition (NER) is an area in which medical named entities are recognized from medical texts, such as diseases, drugs, surgery reports, anatomical parts, and examination documents. Conventional medical NER methods do not make full use of un-labelled medical texts embedded in medical documents. To address this issue, we proposed a medical NER approach based on pre-trained language models and a domain dictionary. First, we constructed a medical entity dictionary by extracting medical entities from labelled medical texts and collecting medical entities from other resources, such as the Yidu-N4K data set. Second, we employed this dictionary to train domain-specific pre-trained language models using un-labelled medical texts. Third, we employed a pseudo labelling mechanism in un-labelled medical texts to automatically annotate texts and create pseudo labels. Fourth, the BiLSTM-CRF sequence tagging model was used to fine-tune the pre-trained language models. Our experiments on the un-labelled medical texts, which were extracted from Chinese electronic medical records, show that the proposed NER approach enables the strict and relaxed F1 scores to be 88.7% and 95.3%, respectively.

Human emotion modeling (HEM): an interface for IoT systems

The use of IoT-based Emotion Recognition (ER) systems is in increasing demand in many domains such as active and assisted living (AAL), health care and industry. Combining the emotion and the context in a unified system could enhance the human support scope, but it is currently a challenging task due to the lack of a common interface that is capable to provide such a combination. In this sense, we aim at providing a novel approach based on a modeling language that can be used even by care-givers or non-experts to model human emotion w.r.t. context for human support services. The proposed modeling approach is based on Domain-Specific Modeling Language (DSML) which helps to integrate different IoT data sources in AAL environment. Consequently, it provides a conceptual support level related to the current emotional states of the observed subject. For the evaluation, we show the evaluation of the well-validated System Usability Score (SUS) to prove that the proposed modeling language achieves high performance in terms of usability and learn-ability metrics. Furthermore, we evaluate the performance at runtime of the model instantiation by measuring the execution time using well-known IoT services.

AVA: A component-oriented abstraction layer for virtual plug&produce automation systems engineering

The prevailing system and software model in automation systems engineering is defined by the IEC 61131 norm. It is to date the best way we know how to express low-level logic and manipulate electrical hardware signals. However, the exponential technological growth is continuing to raise the expectations on what automation systems are supposed to be capable of doing. Fulfilling rising requirements and managing the exploding complexity requires a systematic support for high-level descriptions, structuring, and communication, which the original approach was not built to provide. This work proposes the introduction of an abstraction layer, a component-container infrastructure, defined on top of standard system and software models in automation and mirroring the world of cyber–physical systems, where independent components are interconnected to realize the systems’ purpose by using each other’s functionalities. The concept is implemented in the form of a domain-specific modeling language, applying a classical two-level Model-driven Software Engineering (MDSE) approach. By engineering distinct industrial use cases in accordance with the proposed approach, it is shown that the defined abstractions and mechanisms are capable of expressing the nuances of software design in different domains and can enable the streamlining of the automation systems engineering workflow into a virtual plug&produce process.

Model-driven round-trip engineering for TinyOS-based WSN applications

Wireless Sensor Network (WSN) applications working on TinyOS operating system is widely used in various areas. However, the requirement of managing the power constraints makes TinyOS different from ordinary systems and hence building WSNs with TinyOS can be a challenging and time-consuming task. As successfully applied in many other domains, model-driven engineering (MDE) can facilitate the design and implementation of such applications. Within this context, the researchers have performed noteworthy studies on deriving various MDE approaches and tools. However, these studies do not support the synchronization between TinyOS application models and the generated programs especially when any change is made in the programs. Hence, in this paper, we introduce a model-driven round-trip engineering (RTE) methodology in which both the MDE of TinyOS applications and synchronization between TinyOS instance models and corresponding code are provided with the use of a toolchain. A domain-specific modeling language, called DSML4TinyOS is used for the MDE of such applications while existing TinyOS programs can be reverse engineered in an environment, called RE4TinyOS. Models retrieved from the existing programs can be automatically processed again with DSML4TinyOS. Evaluation results showed that it is possible to obtain the configurations of the TinyOS applications completely just modeling with DSML4TinyOS whereas the same process leads the automatic creation of almost half of the module parts of these applications. Moreover, the time required for developing such systems from scratch decreased approximately to its half. The results also showed that both model-code synchronization and the integration of existing TinyOS applications which do not have system models previously, into the proposed MDE are possible with using RE4TinyOS. RE4TinyOS succeeded in the reverse engineering of all main parts of the TinyOS applications taken from the official TinyOS Github repository and generated models were able to be visually processed in the MDE environment for further modifications.

The Research on Driving Meta-model of Large Rotary-type Cylindrical Antenna

As a complex system with highly integrated multi-discipline, the large rotary-type cylindrical antenna is favored by more and more researchers in IPS detection and large-area survey observation. Due to its complex structure, huge volume and large motion inertia, high requirements on the drive system are indispensable, and there exist some certain limitations in the traditional system design method. This article proposes a large rotary-type cylindrical antenna driving system modeling method based on the meta-model. The domain-specific modeling language of the antenna driving system is defined by extending SysML, and the domain meta-model of the antenna driving system is constructed on this basis. The functional modeling of the antenna driving system is based on the meta-model to support the analysis of the antenna driving process, improve the modeling efficiency and model reusability, and provide a research foundation for the subsequent system modeling design of large antenna.

Quantitative Security Risk Modeling and Analysis with RisQFLan

Domain-specific quantitative modeling and analysis approaches are fundamental in scenarios in which qualitative approaches are inappropriate or unfeasible. In this paper, we present a tool-supported approach to quantitative graph-based security risk modeling and analysis based on attack-defense trees. Our approach is based on QFLan, a successful domain-specific approach to support quantitative modeling and analysis of highly configurable systems, whose domain-specific components have been decoupled to facilitate the instantiation of the QFLan approach in the domain of graph-based security risk modeling and analysis. Our approach incorporates distinctive features from three popular kinds of attack trees, namely enhanced attack trees, capabilities-based attack trees and attack countermeasure trees, into the domain-specific modeling language. The result is a new framework, called RisQFLan, to support quantitative security risk modeling and analysis based on attack-defense diagrams. By offering either exact or statistical verification of probabilistic attack scenarios, RisQFLan constitutes a significant novel contribution to the existing toolsets in that domain. We validate our approach by highlighting the additional features offered by RisQFLan in three illustrative case studies from seminal approaches to graph-based security risk modeling analysis based on attack trees.

A technique for evaluating and improving the semantic transparency of modeling language notations

The notation of a modeling language is of paramount importance for its efficient use and the correct comprehension of created models. A graphical notation, especially for domain-specific modeling languages, should therefore be aligned to the knowledge, beliefs, and expectations of the targeted model users. One quality attributed to notations is their semantic transparency, indicating the extent to which a notation intuitively suggests its meaning to untrained users. Method engineers should thus aim at semantic transparency for realizing intuitively understandable notations. However, notation design is often treated poorly—if at all—in method engineering methodologies. This paper proposes a technique that, based on iterative evaluation and improvement tasks, steers the notation toward semantic transparency. The approach can be efficiently applied to arbitrary modeling languages and allows easy integration into existing modeling language engineering methodologies. We show the feasibility of the technique by reporting on two cycles of Action Design Research including the evaluation and improvement of the semantic transparency of the Process-Goal Alignment modeling language notation. An empirical evaluation comparing the new notation against the initial one shows the effectiveness of the technique.

Towards a Domain-Specific Modeling Language for Extracting Event Logs from ERP Systems

Enterprise resource planning (ERP) systems are often seen as viable sources of data for process mining analysis. To perform most of the existing process mining techniques, it is necessary to obtain a valid event log that is fully compliant with the eXtensible Event Stream (XES) standard. In ERP systems, such event logs are not available as the concept of business activity is missing. Extracting event data from an ERP database is not a trivial task and requires in-depth knowledge of the business processes and underlying data structure. Therefore, domain experts require proper techniques and tools for extracting event data from ERP databases. In this paper, we present the full specification of a domain-specific modeling language for facilitating the extraction of appropriate event data from transactional databases by domain experts. The modeling language has been developed to support complex ambiguous cases when using ERP systems. We demonstrate its applicability using a case study with real data and show that the language includes constructs that enable a domain expert to easily model data of interest in the log extraction step. The language provides sufficient information to extract and transform data from transactional ERP databases to the XES format.

A Modeling Method in Support of Strategic Analysis in the Realm of Enterprise Modeling

Organizations increasingly have to cope with the digital transformation, which is ubiquitous in today’s society. Strategic analysis is an important first step towards the success of digital transformation initiatives, whereby all the elements (e.g., business processes and IT infrastructure) that are required to achieve the transformation can be aligned to the strategic goals and decisions. In this paper, we work towards a modeling method to perform model-based strategic analysis. We explicitly account for information technology (IT) infrastructure because of its key role for digital transformation. Specifically, (1) based on a conducted study on business scholar literature and existing work in conceptual modeling, a set of requirements is first identified; (2) then, we propose a modeling method that integrates, among others, goal modeling, strategic modeling, and IT infrastructure modeling. The method exploits, among others, three previously designed domain specific modeling languages in the Multi-Perspective Enterprise Modeling (MEMO) family: GoalML, SAML and ITML; (3) we illustrate the use of the modeling method in terms of a digital transformation initiative in the electricity sector; and finally, (4) we evaluate the proposed modeling method by comparing it with the conventional SWOT analysis and reflecting upon the fulfillment of the identified requirements.

An Approach of Defining Domain Constraints for Domain-Specific Modeling Language

Many Domain-Specific Modeling Languages (DSML) cannot formally define their semantics, leading to difficulties in identifying user-defined domain constraints. In this study, we propose a user-defined mechanism of domain constraints based on the formalization of structural semantics of DSML. First, we formally define concepts and decision methods of consistency and validity of domain constraints. Subsequently, we establish concepts and reasoning methods of domain-based model consistency. Thus, several domain constraint instances are defined and different models instances’ consistency are reasoned based on formalization of software architecture domain metamodel to illustrate our approach. Finally, our formal definition mechanism of domain constraint is added to our automatic translator for formalizing DSML and its models to automatically reason about domain constraints built based on DSML.

RoBMEX: ROS-based modelling framework for end-users and experts

Autonomous vehicles, such as drones, are gaining great popularity due to their usability and versatility. Nowadays, a significant number of them operate using open source software, such as Robot Operating System (ROS) and the de facto standard MAVLink communication protocol, as they are free and many of them are reusable enough that they can be deployed in various different vehicles. Although these technologies offer a wide variety of resources, using them requires a reasonable background of programming and system engineering. Often, this is not achievable by common drone end-users in the short-term, as they would need to acquire a considerably large amount of know-how before working on specific domains. However, a graphical Domain Specific Modelling Language (DSML) might provide a shortcut to design drone missions using already known concepts to the end-users (or, at least, ones easier to learn). Pursuing this shortcut, RoBMEX is presented as a top-down methodology based on a set of domain specific languages able to enhance the autonomy of ROS-based systems, by allowing the creation of missions graphically, and then generating automatically executable source codes conforming to the designed missions.

Using recommender systems to improve proactive modeling

This article investigates using recommender systems within graphical domain-specific modeling languages (DSMLs). The objective of using recommender systems within a graphical DSML is to overcome a shortcoming of proactive modeling where the modeler must inform the model intelligence engine how to progress when it cannot automatically determine the next modeling action to execute (e.g., add, delete, or edit). To evaluate our objective, we implemented a recommender system into the Proactive Modeling Engine, which is an add-on for the Generic Modeling Environment. We then conducted experiments to subjectively and objectively evaluate enhancements to the Proactive Modeling Engine. The results of our experiments show that extending proactive modeling with a recommender system results in an average reciprocal hit-rank of 0.871. Likewise, the enhancements yield a System Usability Scale rating of 77. Finally, user feedback shows that integrating recommender systems into DSMLs increases usability and learnability.

Blended graphical and textual modelling for UML profiles: A proof-of-concept implementation and experiment

Domain-specific modelling languages defined by extending or constraining the Unified Modelling Language (UML) through the profiling mechanism have historically relied on graphical notations to maximise human understanding and facilitate communication among stakeholders. Other notations, such as text-, form-, or table-based are, however, often preferred for specific modelling purposes, due to the nature of a specific domain or the available tooling, or for personal preference. Currently, the state of the art support for UML-based languages provides an almost completely detached, or even entirely mutually exclusive, use of graphical and textual modelling. This becomes inadequate when dealing with the development of modern systems carried out by heterogeneous stakeholders. Our intuition is that a modelling framework based on seamless blended multi-notations can disclose several benefits, among which: flexible separation of concerns, multi-view modelling based on multiple notations, convenient text-based editing operations (inside and outside the modelling environment), and eventually faster modelling activities.In this paper we report on: (i) a proof-of-concept implementation of a framework for UML and profiles modelling using blended textual and graphical notations, and (ii) an experiment on the framework, which eventually shows that blended multi-notation modelling performs better than standard single-notation modelling.

ALBA: a model-driven framework for the automatic generation of android location-based apps.

In recent years, the number of smartphone users has increased dramatically. These users download millions of apps and use them for various services. Due to the significant demand for mobile apps, developers often seek faster development methods and more effective tools and techniques to generate these apps. Many of these apps are location-based apps in which users receive services based on their geographical location. In this paper, we propose a model-driven approach for the automatic generation of Android location-based mobile apps. Our framework, called ALBA, consists of a domain-specific modeling language, a modeling tool, and a plugin which includes model to code transformations. The modeling tool enables a novice designer to model a location-based app. The model is validated against the predefined constraints and the editor prevents creating invalid models. The designer uses the plugin to generate the Android code of the app. The evaluation of our work is two fold. First, to evaluate the generalizability of the ALBA framework, we conducted an experiment which includes the generation of four industrial location-based apps. Second, to evaluate the usability and quality of both the framework and the generated apps, we conducted a case study consists of three experiments. The results of the evaluation are promising both in terms of the applicability of the framework and the quality of the generated apps.

AgentDSM-Eval: A framework for the evaluation of domain-specific modeling languages for multi-agent systems

Software development required for constructing multi-agent systems (MAS) usually becomes challenging and time-consuming due to the properties of autonomy, distributedness, and openness of these systems in addition to the complicated nature of internal agent behaviors and agent interactions. To facilitate MAS development, the researchers propose various domain-specific modeling languages (DSMLs) by enriching MAS metamodels with a defined syntax and semantics. Although the descriptions of these languages are given in the related studies with the examples of their use, unfortunately, many are not evaluated in terms of either the usability (being hard to learn, understand and use) or the quality of the generated artifacts. Hence, in this paper, we introduce an evaluation framework, called AgentDSM-Eval, with its supporting tool which can be used to evaluate MAS DSMLs systematically according to various quantitative and qualitative aspects of agent software development. The empirical evaluation, presented by the AgentDSM-Eval framework, was successfully applied for one of the well-known MAS DSMLs. The assessment showed that both MAS domain coverage of DSMLs and the agent developers’ adoption of modeling elements can be determined with this framework. Moreover, the tool's quantitative results can assess MAS DSML's performance on the development time and throughput. AgentDSM-Eval also enables the qualitative assessment of MAS DSML features according to novel quality characteristics and measures, which it defines specifically for the MAS domain.