Domain-specific Modeling Research Articles

Development and Use of Domain-specific Learning Theories, Models, and Instruments in Computing Education

Use of theory within a field of research provides the foundation for designing effective research programs and establishing a deeper understanding of the results obtained. This, together with the emergence of domain-specific theory, is often taken as an indicator of the maturity of any research area. This article explores the development and subsequent usage of domain-specific theories and theoretical constructs (TCs) in computing education research (CER). All TCs found in 878 papers published in three major CER publication venues over the period 2005–2020 were identified and assessed to determine the nature and purpose of the constructs found. We focused more closely on areas related to learning, studying, and progression, where our analysis found 80 new TCs that had been developed, based on multiple epistemological perspectives. Several existing frameworks were used to categorize the areas of CER focus in which TCs were found, the methodology by which they were developed, and the nature and purpose of the TCs. A citation analysis was undertaken, with 1,727 citing papers accessed to determine to what extent and in what ways TCs had been used and developed to inform subsequent work, also considering whether these aspects vary according to different focus areas within computing education. We noted which TCs were used most often and least often, and we present several brief case studies that demonstrate progressive development of domain-specific theory. The exploration provides insights into trends in theory development and suggests areas in which further work might be called for. Our findings indicate a general interest in the development of TCs during the period studied, and we show examples of how different approaches to theory development have been used. We present a framework suggesting how strategies for developing new TCs in CER might be structured and discuss the nature of theory development in relation to the field of CER.

On a DSML Domain Server for Fluids Transmission Pipeline Design and Modeling

Domain specific modelling languages (DSMLs) are special purpose languages that have been designed and tailored for specific application domains. With domain specific features user of the language can construct with very familiar notations and get desired outcomes. The focus in this paper is on the functionality of a domain server in a domain specific modelling language for modelling oil and gas pipeline design. A domain server is coupled to a range of pipeline physical components by a pipeline context model, each of the components having a variety of pipeline built units, attributes and values. The domain server stores data received from at least one of the range of the pipeline physical components, the data including values associated with one or more attributes of the components. The domain server derives a model type for at least one phase from the lifecycle of the pipeline design operation for the first instance of the pipeline built units, based on analytics of information stored in the pipeline built units memory and the component attributes storage, where the model type includes a set of attributes for at least one of the range of pipeline physical components. In addition, the domain server generates an orientation for performance operation of an instance including one or more actions from the pipeline built units corresponding to at least one attribute of the set of attributes Keywords: Attribute sets, model type, built units, pipeline context model (PCM)

Predicting Publication of Clinical Trials Using Structured and Unstructured Data: Model Development and Validation Study.

Publication of registered clinical trials is a critical step in the timely dissemination of trial findings. However, a significant proportion of completed clinical trials are never published, motivating the need to analyze the factors behind success or failure to publish. This could inform study design, help regulatory decision-making, and improve resource allocation. It could also enhance our understanding of bias in the publication of trials and publication trends based on the research direction or strength of the findings. Although the publication of clinical trials has been addressed in several descriptive studies at an aggregate level, there is a lack of research on the predictive analysis of a trial's publishability given an individual (planned) clinical trial description. We aimed to conduct a study that combined structured and unstructured features relevant to publication status in a single predictive approach. Established natural language processing techniques as well as recent pretrained language models enabled us to incorporate information from the textual descriptions of clinical trials into a machine learning approach. We were particularly interested in whether and which textual features could improve the classification accuracy for publication outcomes. In this study, we used metadata from ClinicalTrials.gov (a registry of clinical trials) and MEDLINE (a database of academic journal articles) to build a data set of clinical trials (N=76,950) that contained the description of a registered trial and its publication outcome (27,702/76,950, 36% published and 49,248/76,950, 64% unpublished). This is the largest data set of its kind, which we released as part of this work. The publication outcome in the data set was identified from MEDLINE based on clinical trial identifiers. We carried out a descriptive analysis and predicted the publication outcome using 2 approaches: a neural network with a large domain-specific language model and a random forest classifier using a weighted bag-of-words representation of text. First, our analysis of the newly created data set corroborates several findings from the existing literature regarding attributes associated with a higher publication rate. Second, a crucial observation from our predictive modeling was that the addition of textual features (eg, eligibility criteria) offers consistent improvements over using only structured data (F1-score=0.62-0.64 vs F1-score=0.61 without textual features). Both pretrained language models and more basic word-based representations provide high-utility text representations, with no significant empirical difference between the two. Different factors affect the publication of a registered clinical trial. Our approach to predictive modeling combines heterogeneous features, both structured and unstructured. We show that methods from natural language processing can provide effective textual features to enable more accurate prediction of publication success, which has not been explored for this task previously.

The power of transfer learning in agricultural applications: AgriNet.

Advances in deep learning and transfer learning have paved the way for various automation classification tasks in agriculture, including plant diseases, pests, weeds, and plant species detection. However, agriculture automation still faces various challenges, such as the limited size of datasets and the absence of plant-domain-specific pretrained models. Domain specific pretrained models have shown state of art performance in various computer vision tasks including face recognition and medical imaging diagnosis. In this paper, we proposeAgriNetdataset, a collection of 160k agricultural images from more than 19 geographical locations, several images captioning devices, and more than 423 classes of plant species and diseases. We also introduceAgriNetmodels, a set of pretrained models on five ImageNet architectures: VGG16, VGG19, Inception-v3, InceptionResNet-v2, and Xception.AgriNet-VGG19 achieved the highest classification accuracy of 94% and the highest F1-score of 92%. Additionally, all proposed models were found to accurately classify the 423 classes of plant species, diseases, pests, and weeds with a minimum accuracy of 87% for the Inception-v3 model. Finally, experiments to evaluate of superiority ofAgriNetmodels compared to ImageNet models were conducted on two external datasets: pest and plant diseases dataset from Bangladesh and a plant diseases dataset from Kashmir.

A Hybrid Semantic Networks Construction Framework for Engineering Design

Abstract This paper proposes a novel framework for building semantic networks from a seed design statement using Recursive Object Modeling (ROM), Word2Vec language modeling, and vector semantic-based method. Semantic Scholar API was used to retrieve abstracts of scientific papers to build ROM-based Semantic Networks to address the design problem implied in the seed design statement, following Environment Analysis from Environment-Based Design (EBD) methodology. The proposed framework was applied to construct the semantic network for a project to design aircraft braking systems, which demonstrates the framework's efficiency. The presented research makes two major contributions: a ROM-based phrase extractor and a domain-specific language model, which is trained on the automatically collected literature abstracts. Using a manually created and assessed truth set containing 100 pairs of abstract-key phrases, the phrase extractor was evaluated by benchmarking it with two existing off-the-shelf key phrase extraction algorithms: TextRank and Rake. The ROM-based phrase extractor extracted most key phrases from target domains and showed higher precision, recall, and F-1 scores than other methods. Meanwhile, the trained project-specific language model was evaluated using the NASA thesaurus. We randomly sampled 457 pairs of connected domain-specific terms related to aircraft braking and landing knowledge. Our Skip-gram model was compared with Google's pre-trained word2vec model and a baseline word2vec model. The results demonstrated that our language model could detect the most pairs of concepts from the NASA thesaurus. The generated semantic network can be applied to design information retrieval, computer-aided design idea generation, cross-domain communication support system, and designer training tool.

Domain-ensemble learning with cross-domain mixup for thoracic disease classification in unseen domains

Chest radiography is the most common radiology examination for thoracic disease diagnosis, such as pneumonia. A tremendous number of chest X-rays prompt data-driven deep learning models in constructing computer-aided diagnosis (CAD) systems for thoracic disease diagnosis. However, in clinically realistic practice, a trained deep learning model often suffers from performance degradation when applied to unseen data, mainly due to domain shifts caused by various scanner vendors, imaging protocols, and patient demographics etc. To this end, we present a novel domain-ensemble learning with cross-domain mixup (DELCOM) method for thoracic disease diagnosis in unseen domains using chest X-rays. The core idea of our method is to learn multiple domain-specific models using data from both virtual and source domains and to generalize them to unseen domains. Specifically, we first introduce cross-domain mixup to construct virtual samples not belonging to any source domain by mixing up the pairs of samples from source domains. We then model the simulation process of domain shift through domain-ensemble learning. The results indicate that our method not only outperforms other state-of-the-art domain generalization (DG) methods on unseen datasets but also achieves comparable performance across all source domains.

Foundational patterns benchmark

Recently, there has been growing interest in the use of ontology as a fundamental tool for representing domain-specific conceptual models to improve the semantics, accuracy, and relevance of domain users’ query results. Although the amount of data has grown steadily over the past decade, much data shares similar characteristics that can be captured by a foundational ontology. In this paper, we show how queries based on a foundational ontology can be evaluated and their performance measured. We also present a Foundational Patterns benchmark to help select the most efficient triple memory and its layout. We evaluate the foundational benchmark with both generated and real datasets for state-of-the-art triple stores.

Constructing and analyzing domain-specific language model for financial text mining

The application of natural language processing (NLP) to financial fields is advancing with an increase in the number of available financial documents. Transformer-based models such as Bidirectional Encoder Representations from Transformers (BERT) have been successful in NLP in recent years. These cutting-edge models have been adapted to the financial domain by applying financial corpora to existing pre-trained models and by pre-training with the financial corpora from scratch. In Japanese, by contrast, financial terminology cannot be applied from a general vocabulary without further processing. In this study, we construct language models suitable for the financial domain. Furthermore, we compare methods for adapting language models to the financial domain, such as pre-training methods and vocabulary adaptation. We confirm that the adaptation of a pre-training corpus and tokenizer vocabulary based on a corpus of financial text is effective in several downstream financial tasks. No significant difference is observed between pre-training with the financial corpus and continuous pre-training from the general language model with the financial corpus. We have released our source code and pre-trained models.

Machine understanding surgical actions from intervention procedure textbooks

The automatic extraction of procedural surgical knowledge from surgery manuals, academic papers or other high-quality textual resources, is of the utmost importance to develop knowledge-based clinical decision support systems, to automatically execute some procedure’s step or to summarize the procedural information, spread throughout the texts, in a structured form usable as a study resource by medical students. In this work, we propose a first benchmark on extracting detailed surgical actions from available intervention procedure textbooks and papers. We frame the problem as a Semantic Role Labeling task. Exploiting a manually annotated dataset, we apply different Transformer-based information extraction methods. Starting from RoBERTa and BioMedRoBERTa pre-trained language models, we first investigate a zero-shot scenario and compare the obtained results with a full fine-tuning setting. We then introduce a new ad-hoc surgical language model, named SurgicBERTa, pre-trained on a large collection of surgical materials, and we compare it with the previous ones. In the assessment, we explore different dataset splits (one in-domain and two out-of-domain) and we investigate also the effectiveness of the approach in a few-shot learning scenario. Performance is evaluated on three correlated sub-tasks: predicate disambiguation, semantic argument disambiguation and predicate-argument disambiguation. Results show that the fine-tuning of a pre-trained domain-specific language model achieves the highest performance on all splits and on all sub-tasks. All models are publicly released.

Evolution of Domain-Specific Modeling Language: An Example of an Industrial Case Study on an RT-Sequencer

Model-driven engineering is a well-established software development methodology that uses models to develop applications where the end-users with visual elements model abstractions from a specific domain. These models are based on domain-specific modeling language (DSML), which is particular to the problem domain. During DSML use, new ideas emerge and DSMLs evolve. However, reports on DSML evolution are rare. This study presents a new DSML called RT-Sequencer that evolved from our DSML Sequencer to support, in addition to the Data Acquisition domain, also a new domain—Real-Time Control (RTC) systems. The process of defining models with a new language RT-Sequencer has changed in a way that new end-users were introduced—advanced end-users, which use general-purpose language (GPL) and advanced programming concepts to define modeling environments for the RT-Sequencer end-users. More specifically, an industrial experience with the RT-Sequencer is presented, where DSML was opened for extension so that a GPL code could be inserted into the model to create new visual blocks for the end-user, and the possibility to adapt and optimize the execution code for a particular task. Our experience shows the specific case of DSML evolution supporting another problem domain, and the implementation effort needed to extend domain-specific modeling language with GPL support.

Design Ontology Supporting Model-Based Systems Engineering Formalisms

Model-based systems engineering (MBSE) provides an important capability for managing the complexities of system development. MBSE empowers the formalism of system architectures for supporting model-based requirement elicitation, specification, design, development, testing, fielding, etc. However, the modeling languages and techniques are heterogeneous, even within the same enterprise system, which leads to difficulties for data interoperability. The discrepancies among data structures and language syntaxes make information exchange among MBSE models more difficult, resulting in considerable information deviations when connecting data flows across the enterprise. Therefore, this article presents an ontology based upon <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">graphs</i> , <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">objects</i> , <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">points</i> , <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">properties</i> , <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">roles</i> , and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">relationships</i> with <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">extensions</i> (GOPPRRE), providing metamodels that support the various MBSE formalisms across lifecycle stages. In particular, knowledge graph models are developed to support unified model representations to further implement ontological data integration based on GOPPRRE throughout the entire lifecycle. The applicability of the MBSE formalism is verified using quantitative and qualitative approaches. Moreover, the GOPPRRE ontologies are used to create the MBSE formalisms in a domain-specific modeling tool, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">MetaGraph</i> , for evaluating its availability. The results demonstrate that the proposed ontology supports the formal structures and descriptive logic of the systems engineering lifecycle.

Modeling Motivational Interviewing Strategies on an Online Peer-to-Peer Counseling Platform

Millions of people participate in online peer-to-peer support sessions, yet there has been little prior research on systematic psychology-based evaluations of fine-grained peer-counselor behavior in relation to client satisfaction. This paper seeks to bridge this gap by mapping peer-counselor chat-messages to motivational interviewing (MI) techniques. We annotate 14,797 utterances from 734 chat conversations using 17 MI techniques and introduce four new interviewing codes such as ''chit-chat'' and ''inappropriate'' to account for the unique conversational patterns observed on online platforms. We automate the process of labeling peer-counselor responses to MI techniques by fine-tuning large domain-specific language models and then use these automated measures to investigate the behavior of the peer counselors via correlational studies. Specifically, we study the impact of MI techniques on the conversation ratings to investigate the techniques that predict clients' satisfaction with their counseling sessions. When counselors use techniques such as reflection and affirmation, clients are more satisfied. Examining volunteer counselors' change in usage of techniques suggest that counselors learn to use more introduction and open questions as they gain experience. This work provides a deeper understanding of the use of motivational interviewing techniques on peer-to-peer counselor platforms and sheds light on how to build better training programs for volunteer counselors on online platforms.

Composition of multilevel domain-specific modelling languages

Multilevel Modelling (MLM) approaches make it possible for designers and modellers to work with an unlimited number of abstraction levels to specify their domain-specific modelling languages (DSMLs). To fully exploit MLM techniques, we need powerful model composition operators. Indeed, the composition of DSMLs is becoming increasingly relevant to the modelling community either because some DSMLs may share commonalities that we want to make reusable, or because we want to facilitate interoperability between DSMLs. In this paper, we propose a composition mechanism for structure and behaviour of multilevel modelling hierarchies. Our approach facilitates the inclusion of additional features while keeping a clear separation of concerns that enhances modularity. We provide a formal semantics of the constructions based on category theory and graph transformations, and show their use in practice on a case study.

Validation and verification in domain-specific modeling method engineering: an integrated life-cycle view

Validation and verification in domain-specific modeling method engineering: an integrated life-cycle view

An Exploratory Study on Utilising the Web of Linked Data for Product Data Mining

The Linked Open Data practice has led to a significant growth of structured data on the Web. While this has created an unprecedented opportunity for research in the field of Natural Language Processing, there is a lack of systematic studies on how such data can be used to support downstream NLP tasks. This work focuses on the e-commerce domain and explores how we can use such structured data to create language resources for product data mining tasks. To do so, we process billions of structured data points in the form of RDF n-quads, to create multi-million words of product-related corpora that are later used in three different ways for creating language resources: training word-embedding models, continued pre-training of BERT-like language models, and training machine translation models that are used as a proxy to generate product-related keywords. These language resources are then evaluated in three downstream tasks, product classification, linking, and fake review detection using an extensive set of benchmarks. Our results show word embeddings to be the most reliable and consistent method to improve the accuracy on all tasks (with up to 6.9% points in macro-average F1 on some datasets). Contrary to some earlier studies that suggest a rather simple but effective approach such as building domain-specific language models by pre-training using in-domain corpora, our work serves a lesson that adapting these methods to new domains may not be as easy as it seems. We further analyse our datasets and reflect on how our findings can inform future research and practice.

MetaMorph: formalization of domain-specific conceptual modeling methods—an evaluative case study, juxtaposition and empirical assessment

Models have evolved from mere pictures supporting human understanding and communication to sophisticated knowledge structures processable by machines and establish value through their processing capabilities. This entails an inevitable need for computer-understandable modeling languages and causes formalization to be a crucial part in the lifecycle of engineering a modeling method. An appropriate formalism must be a means for providing a structural definition to enable a theoretical investigation of conceptual modeling languages and a unique, unambiguous way of specifying the syntax and semantics of an arbitrary modeling language. For this purpose, it must be generic and open to capturing any domain and any functionality. This paper provides a pervasive description of the formalism MetaMorph based on logic and model theory—an approach fulfilling the requirements above for modeling method engineering. The evaluation of the formalism is presented following three streams of work: First, two evaluative case studies illustrate the applicability of MetaMorph formalism concept by concept on the modeling language ProVis from the domain of stochastic education and the well-known Entity-Relationship language. ProVis as well as ER comprise only a few objects and relation types but with high interconnection and expressive power and are therefore considered interesting specimens for formalization. Second, a comprehensive juxtaposition of MetaMorph to three other formalization approaches based on different foundational theories is outlined concept by concept to underpin the formalism design. Third, an empirical evaluation has been performed, assessing the usability and adequacy of the formalism within a classroom assessment. The results allow for conclusions on the completeness, intuitiveness, and complexity as well as on interdependencies with engineers’ skills.

Critical assessment of transformer-based AI models for German clinical notes

Healthcare data such as clinical notes are primarily recorded in an unstructured manner. If adequately translated into structured data, they can be utilized for health economics and set the groundwork for better individualized patient care. To structure clinical notes, deep-learning methods, particularly transformer-based models like Bidirectional Encoder Representations from Transformers (BERT), have recently received much attention. Currently, biomedical applications are primarily focused on the English language. While general-purpose German-language models such as GermanBERT and GottBERT have been published, adaptations for biomedical data are unavailable. This study evaluated the suitability of existing and novel transformer-based models for the German biomedical and clinical domain. We used 8 transformer-based models and pre-trained 3 new models on a newly generated biomedical corpus, and systematically compared them with each other. We annotated a new dataset of clinical notes and used it with 4 other corpora (BRONCO150, CLEF eHealth 2019 Task 1, GGPONC, and JSynCC) to perform named entity recognition (NER) and document classification tasks. General-purpose language models can be used effectively for biomedical and clinical natural language processing (NLP) tasks, still, our newly trained BioGottBERT model outperformed GottBERT on both clinical NER tasks. However, training new biomedical models from scratch proved ineffective. The domain-adaptation strategy's potential is currently limited due to a lack of pre-training data. Since general-purpose language models are only marginally inferior to domain-specific models, both options are suitable for developing German-language biomedical applications. General-purpose language models perform remarkably well on biomedical and clinical NLP tasks. If larger corpora become available in the future, domain-adapting these models may improve performances.

Extracting interrelated information from road-related social media data

The social media data (SMD) have been viewed as a potential and promising information source of road conditions. However, most existing SMD-based sensing approaches (SMDSAs) either ignore interrelations among information items (e.g., name, direction, and status of the road) or rely on rigid grammar rules to establish entities’ interrelations. Additionally, current SMDSAs in the transportation domain are unable to link the extracted text-formatted information with domain-specific models (e.g., virtual road model, VRM). In order to fill such gaps, this work proposes an improved SMDSA of road conditions, which involves a three-stage (i.e., SMD classification, relation inference, and entity pair recognition) interrelated information extraction model, as well as a semantic converter to feed the SMD-provided text-formatted information into VRMs. The proposed SMDSA is demonstrated by the newly annotated datasets of tweets in Lexington, USA. The three-stage interrelated information extraction model outperforms conventional rule-based methods and deep-learning algorithms (e.g., Text CNN, Bi-LSTM, Piecewise CNN, and Capsule Net). The SMD-enabled VRM also preliminarily shows its capacity to optimize signal timings during incidents that change the road network topology. This work contributes to circumventing the reliance on human-made rules during SMDSAs’ development, bridging user-generated SMD with operable VRMs for potential real-world road management, and providing a standard tweet dataset annotated with interrelation triplets to help promote SMDSA studies.

Production processes modelling within digital product manufacturing in the context of Industry 4.0

Industry 4.0 aims to establish highly flexible production, enabling effective and efficient mass customisation of products. Modelling techniques and simulation of production processes are among the core techniques of the manufacturing industry that facilitate flexibility and automation of a shop floor in the era of Industry 4.0. In this paper, we present an approach to support production process modelling and process model management. The approach is based on Model-Driven (MD) principles and comprises a Domain-Specific Modelling Language (DSML) named Multi-Level Production Process Modelling Language (MultiProLan). MultiProLan uses a set of concepts to specify production process models suitable for automatic instruction generation and execution of the instructions in a simulation or on a shop floor. By using MultiProLan, process designers may create process models independent of the specific production system. Such process models can either be automatically enriched by matching and scheduling algorithms or manually enriched by a process designer via MultiProLan’s modelling tool. In this paper, we also present an application of our approach in the assembly industry to showcase its dynamic resource management, generation of production documentation, error handling and process monitoring.

HealMA: a model-driven framework for automatic generation of IoT-based Android health monitoring applications.

The development of IoT-based Android health monitoring mobile applications (apps) using traditional software development methods is a challenging task. Developers need to be familiar with various programming languages to manage the heterogeneity of hardware and software systems and to support different communication technologies. To address these problems, in this paper, we first analyze the domain of health monitoring mobile applications and then propose a framework based on model-driven engineering that accelerates the development of such systems. The proposed framework, called HealMA, includes a domain-specific modeling language, a graphical modeling editor, several validation rules, and a set of model-to-code transformations, all packed as an Eclipse plugin. We evaluated the framework to assess its applicability in generating various mobile health applications, as well as its impact on software productivity. To this end, four different health monitoring applications have been automatically generated. Then, we evaluated the productivity of software developers by comparing the time and effort it takes to use HealMA compared to a code-centric process. As part of the evaluation, we also evaluated the usability of HealMA-generated apps by conducting a user study. The results show that HealMA is both applicable and beneficial for automatic generation of usable IoT-based Android health monitoring apps.