DE / EN

Bachelor-Seminar

SM 452 für Bachelor­studierende (Wirtschafts­informatik)

Allgemeines

FSS 2022
Verantwortlicher Dozent Prof. Dr. Armin Heinzl
Veranstaltungs­art Seminar
Leistungs­punkte 5 ECTS (WI ab HWS 2013), 4 ECTS (WI bis HWS 2013)
Sprache Englisch
Prüfungs­form und -umfang Seminarpapier (70%), Presentation (20%), Diskussionsbeitrag (10%)
Prüfungs­termin Siehe Infos zur Veranstaltung
Infos für Studierende Registrierung: Bitte beachten Sie unten stehende Informationen!
Timo Himmelsbach, M.Eng.

Timo Himmelsbach, M.Eng.

Ansprech­partner Bachelor-Seminar

Bei Fragen wenden Sie sich bitte an Timo Himmelsbach​​​​​​​.

Infos zur Veranstaltung

  • Kurzbeschreibung

    Digitale Technologien und die ständig wachsenden Datenmengen verändern unser tägliches Leben und die Wirtschaft radikal. Eingebettet in den Kern der Produkte, Abläufe und Strategien vieler Unternehmen, verändern digitale Technologien bestehende Geschäfte in allen Branchen rasant. Neue Markt­angebote, Geschäftsprozesse und Geschäfts­modelle entstehen rund um den Einsatz dieser digitalen Technologien und führen zu digitalen Innovationen1. Die Allgegenwärtigkeit digitaler Technologien verändert unser Verständnis von Informations­systemen (IS) grundlegend, sowohl was ihre Entwicklung, Koordination und Nutzung als auch die Art und Weise, wie wir mit ihnen interagieren, betrifft. An unserem Lehr­stuhl bieten wir ein breites Spektrum an Forschungs­themen in diesem Bereich an, die auch neue digitale Technologien wie künstliche Intelligenz (KI) und maschinelles Lernen (ML) umfassen. Hierbei nehmen wir vor allem nachfolgende Perspektiven ein: Mensch-Computer-Interaktion, System­entwicklung, Wertschöpfung oder Organisation.

    In unserem Seminar untersuchen wir den Einfluss digitaler Technologien auf Individuen und Organisationen. Dabei verknüpfen wir die angebotenen Themen mit unseren laufenden Forschungs­arbeiten, die auf höchstem internationalem Niveau veröffentlicht wurden und werden.

    1. Nambisan, S., Lyytinen, K. & Yoo, Y. Handbook of Digital Innovation. 2–12 (2020) doi:10.4337/9781788119986.00008. 

    Falls Sie mehr über digitale Innovation erfahren möchten, werfen Sie gerne einen Blick auf unseren Master­studien­gang IS 607 und/oder folgende Leseempfehlung:
    Nambisan, S., Lyytinen, K. & Yoo, Y. Handbook of Digital Innovation, (2020), doi:10.4337/9781788119986. 

    Ziel des Seminars

    In diesem Seminar erwerben Sie die Fähigkeit, bestehende Forschungs­arbeiten zu identifizieren, einzuordnen und zu bewerten. Sie werden lernen, eine eigene Forschungs­agenda zu entwickeln sowie diese zu präsentieren und mit den Teilnehmern des Seminars zu diskutieren. Sie machen sich mit verschiedenen Techniken des wissenschaft­lichen Arbeitens und Schreibens vertraut, so dass Sie optimal auf die Konzeption und Anfertigung Ihrer Master­arbeit vorbereitet werden. Wir bieten fünf verschiedene Themen­bereiche an, die hoffentlich Ihr Interesse wecken.

Nach oben

Frühjahr/Sommer 2022

  • Registrierung

    Registrierung erfolgt ausschließlich über das Registrierungs­portal (erreichbar innerhalb des Uni-Netzwerkes oder per VPN). Nur im unten aufgeführten Zeitraum ist eine Registrierung möglich. Dazu bitte das Seminar im Formular auswählen.

    Registrierungs­zeitraum: siehe Termine

    Anforderungen:

    • Kurzes formloses Motivations­schreiben (maximal 1 Seite):
      Bitte wählen Sie ein Thema und begründen Sie Ihre Wahl, z. B. was Sie besonders interessiert und was Sie lernen möchten. Bitte geben Sie auch zwei alternative Themen an.
    • Lebenslauf und Studien­ergebnisse (Notentranskript)

    Es werden weder Registrierungen per E-Mail noch unvollständige Formulare im Registrierungs­tool berücksichtigt.

  • Themen

    Die Studierenden werden gebeten, ein  formloses Motivations­schreiben (maximal 1 Seite) zu verfassen, in dem Sie dieThemenauswahl darlegen und kurz begründen.Dieses Motivations­schreiben stellt neben dem Lebenslauf und dem Notentranskript eine wichtige Referenz für die Seminarzulassung dar.


    A) Human-Computer Interaction
    Human- Computer Interaction, Avatars Adaptable avatars in online games are virtual representations of players in a virtual space. Not only do they enable players to interact with the game, but they also play a central role in computer-mediated communication between players. To represent and express themselves, players often customize their avatars, such as their gender, face, hair, body, apparel accessories or game-relevant items. In their customizations, players can also optimize their avatars to represent an “ideal” version of themselves. In this, avatars can act as a creative platform for identity construction.
    While this identity construction may be beneficial by boosting self-esteem in the virtual environment, it might also cause conflicts between a player’s online and offline identity. This conflict can cause negative effects in the offline world, including gaming addiction, loss of self-esteem or loss of physical relations­hips, amongst others.
    For this seminar paper, the student should conduct a structured literature review to provide an overview of the unintended social and psychological consequences of mismatches between online and offline identity. In addition, the student is expected to outline which theoretical perspectives have been used by previous studies to understand the relations­hip between avatar use and the identified negative consequences.
    Rosa Holtzwart
    Artificial Intelligence, Fairness AI systems already support humans in various decision contexts as they analyze data with high accuracy and discover new patterns in data. With increasing learning capabilities, AI systems can be used to make independent decisions that only need to be monitored by humans in exceptional cases. However, users are often hesitant to accept AI systems if they can obtain a judgement from a human expert instead.
    In this seminar thesis, you will consider users’ fairness perceptions as one reason why human experts’ decisions are preferred over those made by AI systems. You will investigate (a) if consumers perceive AI decisions as more or less fair than decisions made by human experts and (b) if they base their judgment on the same or different fairness criteria for each agent. You will use the Organizational Justice Theory as foundation to structure your literature review and to develop concrete hypotheses on those differences. The seminar thesis will prepare you to conduct an experiment for your master thesis.

    As preparation for the seminar thesis please consider the following reviews:
    • Starke, C., Baleis, J., Keller, B., & Marcinkowski, F. (2021). Fairness Perceptions of Algorithmic Decision-Making: A Systematic Review of the Empirical Literature. arXiv preprint arXiv:2103.12016.
    • Kordzadeh, N., & Ghasemaghaei, M. (2021). Algorithmic bias: review, synthesis, and future research directions. European Journal of Information Systems, 1-22.
    Dr. Ekaterina Jussupow
    B) Exploring Technological Advances of Artificial Intelligence
    Computational Creativity; Machine Learning; Generative Deep Learning; Generative Adversarial Networks; Variational Autoencoder A generative machine learning (ML) model describes how a dataset is generated in terms of a probabilistic model. By sampling from this model, new data can be generated. Generative Adversarial Networks and Variational Autoencoder are examples for such generative models. They can be used to generate any kind of artefact, ranging from visual arts to musical sequences. Thus, they lay ground to support or conduct creative tasks. Using ML methods to conduct creative tasks is studied in the field of computational creativity (CC). Based on a structured literature review, the goal of this seminar paper is to explore and categorize major generative ML methods applicable for CC applications with regards to their creative abilities, i.e., whether the ML method is suitable to be applied for creative tasks. Hereby, it is also essential to carve out how these ML methods can be designed to achieve better creative capabilities. Deborah Mateja
    Explainable AI, hybrid AI, expert-driven decision aids The “black box nature” and lack of reasoning of Machine and Deep Learning applications are critical in high-stakes domains such as healthcare. Therefore, it has accelerated scientific interest in the development of eXplainable AI (XAI) to provide explanations for the cause of its decisions. In a recent trend, research identified synergies in combining data-driven machine learning and domain knowledge-driven semantics to get the best from each. Incorporating radiologist-interpreted features into learning, for instance, brings a deeper level of understanding of CT images to intelligent systems and enhances interpretation for radiologists in turn. Therefore, this seminar paper assesses the current state of knowledge-informed AI to provide an overview of how recent works incorporate intuitive semantic features in hybrid systems and how these approaches reflect expert-driven decision aids, contributing to explainability. Luis Oberste
    Requirements Elicitation in Information Systems A target-oriented Requirements Engineering (RE), the process of deriving the necessary requirements for a piece of software, is one of the key success factors of Information Systems Development (ISD). Information systems (IS) research has been addressing the topic from a behavioral and empirical as well as from a design science perspective.
    A tremendous evolutionary step in RE is called data-driven RE. Data-driven RE enriches the methods of collecting and analyzing user input in traditional RE with the automated and continuous analysis of novel feedback sources as well as with the analysis of context-aware usage data to identify, prioritize, document, and manage requirements for a software product.
    For this seminar thesis, the student should review current IS and software engineering literature to gain an overview on the main ideas for the second part of data-driven RE, namely leveraging usage-data for RE. The discussion of the findings should compare the identified approaches from a self-derived framework or taxonomy.
    Philipp Hoffmann
    C) Management and Impact of Artificial Intelligence
    Artificial Intelligence, Machine Learning, Knowledge, Business Value of IT Artificial Intelligence (AI) has made stunning progress over the past decade based on big data, scalable and affordable computing power, and increasingly powerful algorithms. State-of-the-art AI relies on Machine Learning (ML) algorithms that can perform sensing, reasoning, and interaction activities without pre-defined solution algorithms, but by learning from patterns in data. Hence, ML systems can be applied to substitute or complement the knowledge work of human professionals. In addition to direct performance effects in organizations (such as increased process efficiency or decision quality), the application of ML systems can also affect organizational knowledge.
    Research has started to discuss how ML systems can create, extend, alter, substitute or displace knowledge at multiple levels of an organization. Recent qualitative, design-oriented, and conceptual studies provide deep insights into the effects of ML systems on different types of organizational knowledge (incl. domain expertise and data science knowledge) held by different actors and groups in different parts of the organization (e.g., system users, developers, domain experts, managers, customers). Nevertheless, a structured overview to outline the impact of ML systems on organizational knowledge is yet missing.
    In this seminar thesis, you will conduct a literature review with the following goals: (1) apply an according to an established theory or framework to summarize existing knowledge regarding the effects of ML systems on organizational knowledge; (2) identify research gaps regarding the effects of ML systems on organizational knowledge and formulate directions for future research.

    Recommended starting points:
    • Nonaka, I. (1994). A dynamic theory of organizational knowledge creation. Organization science, 5(1), 14-37.
    • Alavi, M., & Leidner, D. E. (2001). Knowledge management and knowledge management systems: Conceptual foundations and research issues. MIS Quarterly, 107-136.
    • Recent special issues on “Managing AI” in MISQ (September 2021) and on “Artificial Intelligence in Organizations” in JAIS/MISQE (December 2020)
    • Further information on the scope of the literature review will be provided by the supervisor
    Pascal Kunz
    D) Value Creation in Platform Ecosystems
    Platform ecosystems, platform governance, power distribution, decentralized governance Many of today’s most valuable firms, such as Apple, Facebook, Amazon or SAP, are digital platform owners. In platform strategies, a provider of a core technology (e.g., Google Android) harness outside innovation by involving independent actors (e.g., external app developers) in the firm’s value creation. Unlike traditional relations­hips between seller and buyer, value creation in platforms is organized within an innovative ecosystem of loosely-coupled parties. The platform owner’s activities shift towards coordinating ecosystem participants. The high degree of centralization of existing platform systems in terms of the platform’s governance authority increasingly raises concerns about the asymmetric distribution of power between platform owner and complementors. Decentralized governance (such as in emerging blockchain platforms) promise to address existing shortcomings and concerns regarding centralized governance. In this seminar project, the student is expected to review current research on platform governance to identify benefits and shortcomings of (de-)centralized governance, to elaborate on practical examples from pre-defined contexts and to propose promising directions for future research on platform governance. André Halckenhäußer
    Decentralized Platform Ecosystems New technologies and the increasing amount of data are transforming traditional businesses. Today, the most valuable companies are built on digital platforms that bring together two or more market actors and grow through network effects (1). Besides facilitating matchmaking and the disruptive potential that centralized platform-based businesses come along with, these business models have recently faced increasing resistance from regulators. Increasing market power, unfair competition based on existing data monopoles and a closed innovation frame built a basis for winner-takes-all dynamics and are major challenges regulators are dealing with.
    Yet, platform ecosystems based on distributed ledger technologies (DLT) are promising to have substantial impact on current centralized platform-based businesses by taking out the “middle-man” and providing an open, decentralized environment based on standards that support data interoperability, data sovereignty, as well as security and open innovation.
    Research on decentralized DLT-based platform ecosystems is still in its infancy and lacks insightful studies. In this seminar thesis, the student is expected to conduct a structured literature review in order to elaborate key characteristics of decentralized DLT-based platform ecosystems and the similarities and differences of these emerging business model. In addition, students are expected to elaborate the potential of DLT-based platform ecosystems based on an analysis of two use cases. The use case selection will be discussed with the supervisor.

    • Cusumano, M. A., Yoffie, D. B. & Gawer, A. The Future of Platforms. MIT Sloan Management Review (2020).

    Timo Himmelsbach
    E) Healthcare IT
    mHealth, behavior change, Just-in-time adaptive interventions (JITAIs) An increasing number of individuals use mobile health applications (mHealth) to adopt healthy behaviors and improve health outcomes. However, mHealth use is often limited to few initial interactions and decreases over time, which limits its ability to change individuals’ health behaviors. To overcome this, app developers often nudge mHealth use by sending push notifications. mHealth apps are constantly collecting data on users, their environment, and their behavior (e.g. location, weather, current app use, behavior, and mood), which makes it possible to deliver messages that adapt to an individual’s changing status over time. However, little is known about what content such messages should contain, under what circumstances messages should be sent, and when the best time is to nudge mHealth use and behavior. Just-in-time adaptive interventions (JITAIs) can overcome this limitation by adapting the provision of support (e.g., the type, timing, intensity) over time to an individual's changing status and contexts, with the goal to deliver support at the moment and in the context that the person needs it most and is most likely to be receptive. This thesis will review academic literature that employs JITAIs. The thesis will identify what theoretical perspectives are used, identify research gaps, and provide avenues for future research. The results of the literature review will have implications for how to better design messages to nudge mHealth use and subsequent behavior. In order to do this, this seminar paper will draw on academic literature related to information systems and behavioral science. Monica Fallon
  • Kursüberblick und -termine

    Event Zeitraum / DeadlineArbeits­ergebnisse
    Registrierungs­zeitraum 1.1. – 14.2.2022 (23:59) – Registrierung über das Online-Tool
    – Fügen Sie Ihren Lebenslauf, das Notentranskript und Ihr Motivations­schreiben an
    Versand der Bestätigungen 16.2.2022 (mittags)
     
    Deadline zum Rücktritt 17.2.2022 (mittags)  
    Kick-Off Meeting 24.2.2022
    10:15 - 11:15 Uhr
    Raum: SO 318 (ZOOM-Lehre-099)
    – Teilnahme an der Kick-Off-Einführungs­veranstaltung
    – Kontakt und Treffen mit Ihrem Betreuer
    1. Meilenstein 10.3.2022 Ersten Entwurf bei Ihrem Betreuer einreichen:
    – Detaillierte Gliederung
    – Literatur­verzeichnis
    2. Meilenstein 7.4.2022 Zweiten Entwurf bei Ihrem Betreuer einreichen:
    – Inhaltsverzeichnis
    – Einführung: vollständig formuliert
    – Methodik: vollständig formuliert
    – Ergebnisse: strukturierter Entwurf
    – Diskussion: strukturierter Entwurf
    Abgabe der Arbeit 21.4.2022 (mittags) – Zwei Ausdrucke der Seminararbeit beim Sekretariat einreichen (bis 12 Uhr)
    – Senden Sie eine digitale Version der Seminararbeit per E-Mail an Timo Himmelsbach (himmelsbach uni-mannheim.de) und an Ihren Betreuer (bis 12 Uhr)

    Änderung aufgrund von Corona: Bitte senden Sie Ihre Arbeit bis 12 Uhr im PDF-Format per Mail an Timo Himmelsbach (himmelsbach uni-mannheim.de) und setzen das Sekretariat (wifo1 uni-mannheim.de) auf CC. Bitte reichen Sie zudem, so bald wie möglich nach dem Abgabetermin, zwei Ausdrucke im Sekretariat ein.
    Abgabe der Präsentation 9.5.2022 (mittags) – Optional: Bitten Sie Ihren Betreuer vorab um Feedback zur Präsentation
    – Senden Sie Ihre Präsentation im PDF-Format per E-Mail an Timo Himmelsbach (himmelsbach uni-mannheim.de)
    Präsentation Donnerstag, 12.05.2022 (nachmittags) und Freitag, 13.05.2022 (morgens) – Besuchen Sie das Seminar und beteiligen Sie sich aktiv an der Diskussion am Seminartag
    – Präsentieren und diskutieren Sie Ihre Seminararbeit im gemeinsamen Workshop
    – Diskussion und Feedback für mindestens eine Seminararbeit der anderen Studierenden
  • Literatur