If language shapes our reality, changing our language might lead to a different, potentially preferable reality. This thought is echoed throughout a variety of philosophical schools and can, in different variations, with different assumptions, and with different implications, be found in the writings of Ludwig Wittgenstein, Richard Rorty, Friedrich Nietzsche, Immanuel Kant, or Humberto Maturana. Recently, the discussion has received more widespread attention. Motivated in part from feminist philosophy of the 1990s, philosophers have combined their research efforts towards the improvement of language under the moniker of conceptual engineering and conceptual ethics. The amelioration of concepts and language is faced with several theoretical and practical challenges. What makes a concept better than another? How could a new concept be adopted by respective language users?

Information systems development is essentially concerned with language development (clarification and sources per request). Broadly, this poses the question if information systems can support conceptual engineering and, if so, in what regards. Machine learning (ML) might be a fruitful first step to guide this analysis. Contemporary ML approaches are inductive (cf. Rescher 1980): they generate potentially novel generalizations based on a set of observations. Researchers like Rees (2022) therefore suggest that they might guide the development of novel concepts.

This master’s thesis should explore the capabilities of ML to support conceptual engineering. You should identify potential tasks of conceptual engineering and what requirements they face. Then you should investigate how different ML approaches (we can disucss which in our first meetings) can serve to address these requirements.

Literature

• Burgess A, Cappelen H, Plunkett D (eds) (2020) Conceptual Engineering and Conceptual Ethics. Oxford University Press: Oxford
• Butlin P (2021) Sharing Our Concepts with Machines. Erkenntnis
• Cappelen H, Dever J (2019) Bad Language. Oxford University Press: Oxford
• Haslanger S (2012) Resisting Reality: Social Construction and Social Critique. Oxford University Press: Oxford
• Medin DL, Smith EE (1984) Concepts and Concept Formation. Annual Review of Psychology 35(35):113–138
• Montemayor C (2021) Language and Intelligence. Minds and Machines 31:471–486
• Ontañón S, Dellunde P, Godo L, Plaza E (2012) A Defeasible Reasoning Model of Inductive Concept Learning from Examples and Communication. Artificial Intelligence 193:129–148
• Rees T (2022) Non-Human Words: On GPT-3 as a Philosophical Library. Daedalus 151(2):168–182
• Rescher N (1980) Induction: An Essay on the Justification of Inductive Reasoning. Basil Blackwell: Oxford

Artificial neural networks (ANNs) denote a popular class of models used within machine learning. An ANN typically consists of multiple layers of simple processing units, so-called artificial neurons. Most current ANNs involve multiple layers of these processing units, hence the term deep learning is sometimes applied to describe them. Historically, they emerged from a neurophysiological inspiration to express the processing of mammal neurons in mathematical terms (cf. McCulloch and Pitts 1943). There exists a plethora of different approaches to the design of ANNs, some variations include the number of artificial neurons in a layer, the activation function applied, or the connection of artificial neurons between layers. From these variations have emerged several classes of ANN architectures, such as Multi-Layered Perceptrons (MLPs), Generative Adversial Networks (GANs), Convolutional Neural Networks (CNNs), or Recurrent Neural Networks (RNNs). It is conspicuous many papers, which discuss a particular ANN architecture,represent them in some diagrammatic form. This diagrammatic representation, however, does not follow any unified structure. This results in two challenges: First, ANNs are not visually comparable through an analysis of their diagrammatic representations. Second, the depicted diagrams of ANNs might lack relevant information, overseen by the original researchers. In short: It appears that the depiction of ANNs lack a conceptual modeling language.

The present thesis should adress this gap. Therefore, it is relevant to expound on the foundations and variations of ANNs as well as to explore the fundamentals of conceptual modeling languages. Based on an analysis of the design, evaluation, and application of ANNs, requirements for a corresponding modeling method should be derived. Thereupon, these insights should be used to specify a conceptual modeling method for ANNs.

Introductory Literature:

• Aggarwal CC (2018) Neural Networks and Deep Learning: A Textbook. Springer International Publishing: Cham
• Du K-L, Swamy MNS (2014) Neural Networks and Statistical Learning. Springer-Verlag: London
• Frank U (2013) Domain-Specific Modeling Languages – Requirements Analysis and Design Guidelines. In: Reinhartz-Berger I, Sturm A, Clark T, Wand Y, Cohen S, Bettin J (eds.) Domain Engineering: Product Lines, Conceptual Models, and Languages. Springer: Cham, pp. 133-157
• Kelleher JD (2019) Deep Learning. The MIT Press: Cambridge, MA, London
• McCulloch WS, Pitts W (1943) A Logical Calculus of the Ideas Immanent in Nervous Activity. Bulletin of Mathematical Biophysics 5:115-133

Motivation

Reference models are conceptual models which focus commonalities of organizations, within or across industries and/or application domains (Thomas, 2005) . Reference models hold several promises, like capitalizing on already encoded domain expertise instead of developing a domain model from scratch. As a result, reference models remain a topic of active interest for both practice and academia, as reflected in relatively recent industry reference models like the NISTIR 7628 for smart grid cybersecurity (NIST Smart Grid Cybersecurity Panel, 2010).

However, the creation of reference models comes with several challenges.

As pointed out in de Kinderen & Kaczmarek-Heß (2019) these challenges partly can be traced to the language architecture underlying reference models. In particular these language architectures often do not treat abstraction as a first class citizen. As a result, to model reference models with multiple levels of abstraction, which is typical for industry reference models like the NISTIR 7628, requires (i) workarounds, and (ii) in general comes with a set of inherent limitations which inhibit capitalizing on the full potential of reference modeling.

Objective

The purpose of this master thesis is to make a systematic comparison between conventional meta modeling approaches and multi-level modeling approaches for the purposes of reference modeling. To this end, based upon a motivated selection of a conventional and multi-leveling approach of choice for a well-documented reference model you draw a systematic comparison – specifically as this pertains the specificities of reference models.

As a point of departure, initially the focus will be on NISTIR 7628, which is a well-established reference model for smart grid cybersecurity. However, in discussion with the supervisor, another reference model can also be selected.

Einstiegsliteratur:

U. Frank, “Multilevel modeling – toward a new paradigm of conceptual modeling and information systems design,” BISE, vol. 6, no. 6, pp. 319–337, 2014.

de Kinderen, S., & Kaczmarek-Heß, M. (2019, July). Multi-level Modeling as a Language Architecture for Reference Models: On the Example of the Smart Grid Domain. In 2019 IEEE 21st Conference on Business Informatics (CBI) (Vol. 1, pp. 174-183). IEEE.

NIST Smart Grid Cybersecurity Panel: NISTIR 7628-guidelines for smart grid cyber security vol. 1-3 (2010)

Fettke, P., Loos, P.: Perspectives on reference modeling. In Fettke, P., Loos, P., eds.: Reference Modeling for Business Systems Analysis. (2007) 1-21

Thomas, O.: Understanding the term reference model in information systems research: history, literature analysis and explanation. In: International Conference on Business Process Management, Springer (2005) 484-496

U. Frank, “Evaluation of reference models,” in Reference modeling for business systems analysis. IGI Global, 2007, pp. 118–140.

In der Physik wird schon seit geraumer Zeit die Idee, eine neue, auf der Quantenmechanik basierende Rechnerarchitektur zu entwickeln, diskutiert. Die technische Realisierung ist jedoch mit großen Herausforderungen verbunden. Erst in jüngster Zeit sind Anzeichen dafür erkennbar, dass Quantenrechner in absehbarer Zukunft verfügbar sein werden, so dass sich für eine wachsende Zahl von Unternehmen die Frage stellt, ob und in welcher Form diese Technologie für sie einsetzbar ist und welche Wettbewerbsvorteile sich damit realisieren lassen. In dieser Masterarbeit soll die prinzipielle Funktionsweise von Quantenrechnern untersucht werden. Dazu ist eine Auseinandersetzung mit den physikalischen Grundlagen erforderlich. Darauf aufbauend sollen die wesentlichen Verheißungen dieser Technologie begründet dargestellt werden. Anschließend ist zu untersuchen, für welche Einsatzszenarien in Unternehmen Quantenrechner besonders attraktiv erscheinen.

In practice and academia a wide variety of methods for different classes of typical business informatics problems exists, developed by consultancy agencies, standardization bodies, software companies, universities, and public-private partnerships. Such methods include agile software development methods (Qumer, 2008), or methods for enterprise architecture design and management (Winter et al, 2010). The wide variety of methods is not surprising, as in addressing a particular problem (e.g., developing a new piece of software) the adoption of an appropriate method allows one to, roughly speaking, capitalize on codified experiences with similar problems encountered in the past. Thus, one does not have to start from scratch.  Yet, to be effective the method often needs to be adapted to the context of use (Henderson-Sellers et al (2011)). Here, a major challenge is that there is no uniform view of what are the constituents of a method. For one, whereas agile software development methods are characterized by adherence to a set of principles (from the agile manifesto (Fowler et al, 2001)), enterprise architecture methods typically aim at offering a set of different perspectives (sometimes also called layers, or dimensions (cf. Winter et al, 2006)), with concepts that are relevant for describing each perspective. This hints at what actually characterizes a method in practice can be quite diverse, in terms of adhering to a common way of thinking, providing a process,  a process and a modeling language, or all of the above complemented with heuristics and guidelines. In the end, this lack of a uniform method conception inhibits the systematic adaptation of methods to a given enterprise’s context of use.

The aim of this master’s thesis is to construct a method for constructing methods - i.e., a meta-method. This meta-method shall at the very least consist of a meta-model which describes the core elements of a method and a process, and corresponding guidelines, for instantiating the method, so that the meta-method at hand can be used to create or adapt a method to the situation at hand. To showcase the meta-methods fitness for use, you use it for the construction of a method for a domain/context of choice.

Einstiegsliteratur

Strecker, Stefan, David Heise, and Ulrich Frank. "RiskM: A multi-perspective modeling method for IT risk assessment." Information Systems Frontiers 13.4 (2011): 595-611.

Henderson-Sellers, Brian, and Jolita Ralyté. "Situational method engineering: state-of-the-art review." Journal of Universal Computer Science (2010).

Qumer, Asif, and Brian Henderson-Sellers. "An evaluation of the degree of agility in six agile methods and its applicability for method engineering." Information and software technology 50.4 (2008): 280-295.

Winter, Katharina, et al. "Investigating the State-of-the-Art in Enterprise Architecture Management Methods in literature and Practice." MCIS (2010).

Winter, Robert, and Ronny Fischer. "Essential layers, artifacts, and dependencies of enterprise architecture." Enterprise Distributed Object Computing Conference Workshops, 2006. EDOCW'06. 10th IEEE International. IEEE, 2006.

Fowler, Martin, and Jim Highsmith. "The agile manifesto." Software Development 9.8 (2001): 28-35.

A method aims at solving a class of problems, and typically consists of a linguistic artifact and a process for using it (Strecker et al, 2010). An example following this definition is RiskM (Strecker et al, 2010), a method for IT risk assessment which consists of a conceptual modeling language called RiskM, and a process model plus various heuristics.

In practice a wide variety of methods for different classes of typical business informatics problems exists, developed by consultancy agencies, standardization bodies, and software companies. Such methods include agile software development methods (Qumer, 2008), or methods for enterprise architecture design and management (Winter et al, 2010). The wide variety of methods is not surprising, as in addressing a particular problem (e.g., developing a new piece of software) the adoption of an appropriate method allows one to, roughly speaking, capitalize on codified experiences with similar problems encountered in the past. Thus, one does not have to start from scratch. However, the effectiveness of a method to a large extent depends on how well it fits with the context of use. For one, while there is currently a hype to develop all software in an agile manner, anecdotal remarks from software developers in the banking industry hint that agile methods do not lend themselves well to the nature of the banking domain, as often concerns such as security and stability of a system take prevalence over its quick “agile” development.

Furthermore, even if there is a fit of a method with the domain of use, to be effective the method often needs to be adapted to the context of use (Henderson-Sellers et al (2011)). Here, a major challenge is that there is no uniform view of what are the constituents of a method. For one, whereas agile software development methods are characterized by adherence to a set of principles (from the agile manifesto (Fowler et al, 2001)), enterprise architecture methods typically aim at offering a set of different perspectives (sometimes also called layers, or dimensions (cf. Winter et al, 2006)), with concepts that are relevant for describing each perspective. This hints at what actually characterizes a method in practice can be quite diverse, in terms of adhering to a common way of thinking, providing a process,  a process and a modeling language, or all of the above complemented with heuristics and guidelines. In the end, this lack of a uniform method conception inhibits the systematic adaptation of methods to a given enterprise's context of use.

The aim of this bachelor thesis is to describe and compare methods. In particular, your aim is to observe similarities and differences between the methods being employed, so as to provide a first step towards a common description of what are the core constituencies of a method. As a point of departure for this comparison, one can use ideas from method engineering (e.g., Henderson-Sellers et al (2011)) and the understanding of a method as put forward in Strecker et al. (2010). Note that while you have a choice as to what methods are selected, the criteria for method selection must be grounded in a solid argumentation.

Einstiegsliteratur

Strecker, Stefan, David Heise, and Ulrich Frank. "RiskM: A multi-perspective modeling method for IT risk assessment." Information Systems Frontiers 13.4 (2011): 595-611.

Henderson-Sellers, Brian, and Jolita Ralyté. "Situational method engineering: state-of-the-art review." Journal of Universal Computer Science (2010).

Qumer, Asif, and Brian Henderson-Sellers. "An evaluation of the degree of agility in six agile methods and its applicability for method engineering." Information and software technology 50.4 (2008): 280-295.

Winter, Katharina, et al. "Investigating the State-of-the-Art in Enterprise Architecture Management Methods in literature and Practice." MCIS (2010).

Winter, Robert, and Ronny Fischer. "Essential layers, artifacts, and dependencies of enterprise architecture." Enterprise Distributed Object Computing Conference Workshops, 2006. EDOCW'06. 10th IEEE International. IEEE, 2006.

Fowler, Martin, and Jim Highsmith. "The agile manifesto." Software Development 9.8 (2001): 28-35.

Motivation:

The preparation of course material is a laborious task. Often, it is aimed at the creation of a sequence of slides. Such an approach, however, is unsatisfactory for various reasons. It does not account for the fact that different students may need different course materials. It is restricted to a sequence of pages including text or graphics (videos are possible, too, but are still rarely used). Furthermore, reuse is mainly restricted to copy&paste. As a consequence, this approach is not suited to create material for supporting more efficient ways of teaching and learning. In addition to that, the effort for maintaining course material is extensive.

This thesis aims at the development of a modelling method that fosters the design and management of material for university courses. For this purpose, new, individualized and interactive forms for course material need to be analyzed in order to identify relevant requirements the targeted method should satisfy. The method itself consists of a domain-specific modelling language (DSML) and a corresponding process model. In an ideal case, the DSML will be implemented with a meta-modelling tool.

Motivation:

Unternehmensmodelle müssen persistent gemacht werden. Im einfachsten Fall werden sie dazu in Dateien abgelegt. Eine solche Repräsentation ist allerdings mit erheblichen Nachteilen verbunden. Spezielle Repositories bieten eine bessere Unterstützung der Modellintegrität und erlauben zudem u.U. einen parallelen Zugriff auf Modelldaten. In der Arbeit sind zunächst Anforderungen an Repositories zur Repräsentation von Modellen zu analysieren. Anschließend sind auf dieser Grundlage ausgewählte Modellrepositories vergleichend zu bewerten.

Motivation:

Die Integration von Anwendungssystemen mit korrespondierenden Unternehmensmodellen verspricht eine überaus attraktive Perspektive für die Gestaltung zukünftiger Unternehmenssoftware. Eine entsprechende Vision wurde in einem Forschungsbericht (Frank, Strecker: Beyond ERP Systems: An Outline of Self-Referential Enterprise Systems. ICB-Research Report No. 31, 2009) skizziert. Die Realisierung selbstreferentieller Unternehmenssoftware sieht sich einer Reihe von Herausforderungen gegenüber. Eine besondere Schwierigkeit ist darin zu sehen, dass die Implementierung von Modellierungswerkzeugen und Anwendungssystemen auf unterschiedlichen Abstraktionsebenen stattfindet. Um vom Modell zum Code zu gelangen, ist deshalb die Generierung von Code erforderlich. Dies führt zu dem bekannten Problem Code und Modell konsistent zu halten. Die wesentliche Ursache für das Problem ist darin zu sehen, dass gängige Programmiersprachen nur zwei Abstraktionsebenen - Typ und Instanz - aufweisen. Seit einiger Zeit gewinnen (Meta-) Programmiersprachen an Bedeutung, die mehrere Abstraktionsebenen aufweisen. Eine dieser Sprachen ist XMF. Die Arbeit ist darauf gerichtet, die Eignung der Sprache XMF für die Synchronisation von Code und Modell zu untersuchen.

Motivation:

Die Pflege und Weiterentwicklung von Modellen führt im Zeitverlauf zu verschiedenen Modellvarianten und -versionen. Um eine konsistente Evolution und Variation von Modellen zu unterstützen, sind differenzierte Konzepte von Modellversion und Modellvariante erforderlich. Sie erlauben es, nur das Delta zwischen verschiedenen Versionen bzw. zwischen einem Modell und seinen Varianten zu erfassen und zu speichern. Die Definition entsprechender Konzepte ist mit erheblichen Herausforderungen verbunden. In der Arbeit sind zunächst Vorschläge aus der Literatur zu erfassen udn vergleichend zu evaluieren. Anschließend sind für ausgewählte Modellarten aus dem Kontext der Unternehmensnmodellierung geeignete Konzepte für Modellversionen und -varianten zu spezifizieren.

Motivation:

Multiperspektivische Unternehmensmodelle sind darauf gerichtet, verschiedenen Zielgruppen im Unternehmen eine aufgabengerechte Abbildung ihrer Sicht auf das Unternehmen zu bieten und gleichzeitig einen Beitrag zur Integration dieser Sichten zu leisten. Domänenspezifische Modellierungssprachen, die u.a. eine weitgehend selbstsprechende Notation bieten sollten, sollen dabei die Verständlichkeit von Modellen fördern, den Benutzern im besten Fall einen intuitiven Zugang ermöglichen. Dieses Ziel kann allerdings nicht immer in befriedigender Weise erreicht werden. Zum einen stellen Modelle eine Abstraktion dar, die bewusst auch solche Aspekte des Kontextes ausblenden, die für ein differenziertes Verständnis von Bedeutung sind. Zum anderen mag die Repräsentation semi-formaler Modelle auf manche Akteure nicht ansprechend wirken. Deshalb ist es sinnvoll, Modelle mit natürlichsprachlichen Dokumenten zu annotieren. Allerdings sind die Erstellung und vor allem die Pflege mit einem erheblichen Aufwand verbunden. In dieser Master-Arbeit soll untersucht werden, wie die Erstellung und Pflege von natürlichsprachlichen Annotationen durch geeignete Software unterstützt werden kann. Dazu soll aufgezeigt werden, wie eine existierende Modellierungsumgebung um eine entsprechende Komponente ergänzt werden kann. Es ist zudem möglich, eine prototypische Implementierung zu erstellen.

Einstiegsliteratur:

• Clark, T., Sammut, P., Willans, J.: Applied metamodelling: a foundation for language driven development. 2nd edn., Ceteva 2008, (Online verfügbar)
  Frank, U.: Multi-Perspective Enterprise Modeling: Foundational Concepts, Prospects and Future Research Challenges. In: Software and Systems Modeling, 2013
  Leopold, H.; Mendling, J.; Polyvyanyy, A.: Generating Natural Language Texts from Business Process Models: Advanced Information Systems Engineering Lecture Notes in Computer Science Volume 7328, 2012, pp 64-79

Motivation:

Unternehmensmodelle waren zunächst vor allem darauf gerichtet, eine an den Unternehmenszielen ausgerichtete Gestaltung von Informationssystem und organisatorischem Handlungssystem zu unterstützen. Seit einigen Jahren wird zudem gefordert, Unternehmensmodelle auch zur aktiven Unterstützung des Managements einzusetzen – „Enterprise Models at Runtime“. Dazu wurden Architekturen entwickelt, die eine Integration von Modellierungsumgebungen mit Unternehmenssoftware ermöglichen. Geschäftsmodelle zielen demgegenüber darauf, wesentliche Erfolgsfaktoren eines Unternehmens in eher groben Zusammenhängen so darzustellen, dass sich daraus Schlussfolgerungen auf die Profitabilität ableiten lassen. Sie werden typischerweise vor der Gründung eines Unternehmens erstellt und richten sich nicht zuletzt an Investoren. In dieser Master-Arbeit sollen Unterschiede und Gemeinsamkeiten von Geschäftsmodellen und Unternehmensmodellen herausgearbeitet werden. Vor diesem Hintergrund soll dann am Beispiel zweier ausgewählter Ansätze aufgezeigt werden, wie eine Integration durchgeführt werden könnten und welche Synergieeffekte dabei zu erwarten sind. Dabei soll auch die Nutzung entsprechender Modelle und Werkzeuge in späteren Phasen eines Unternehmenslebenszyklus betrachtet werden.

Einstiegsliteratur:

• Frank, U.: Enterprise Modelling: The Next Steps. In: Enterprise Modelling and Information Systems Architectures. Vol. 9, No. 1, 2014, pp. 22–37
• Frank, U.; Strecker, S.: Beyond ERP Systems: An Outline of Self-Referential Enterprise Systems. ICB-Research Report, Institut für Informatik und Wirtschaftsinformatik (ICB), Universität Duisburg-Essen, No. 31, Universität Duisburg-Essen 2009OMG: Business Motivation Model. Version 1.2 (Online verfügbar)
• Osterwalder, A.; Pigneur, Y.: Business model generation. A handbook for visionaries, game changers, and challengers. Willowbrook, IL 2010