Logo des Repositoriums

Auflistung nach Schlagwort "Natural Language Generation"

1 - 4 von 4
  • Konferenzbeitrag
    Generating Explanations for Algorithmic Decisions of Usage-Based Insurances using Natural Language Generation
    (Software Engineering und Software Management 2018, 2018) Braun, Daniel; Matthes, Florian
    Usage-based insurances are becoming more and more popular, especially for cars. These so called telematics insurances use different sensors installed in a car to track the individual driving style of the driver. Instead of calculating insurance premiums based on statistical risk groups, insurance companies can use these data to create individual risk profiles and calculate insurance premiums accordingly. We present an approach to use Natural Language Generation (NLG) in order to explain customers which aspects of their behaviour influenced the assessment of the algorithm. In this way, we can not only increase the acceptance of customers regarding such systems, but also positively influence their future behaviour.
  • Zeitschriftenartikel
    KI schlägt Mensch?!
    (HMD Praxis der Wirtschaftsinformatik: Vol. 58, No. 4, 2021) Theobald, Elke; Malthaner, Maren; Föhl, Ulrich
    Die wachsende Bedeutung des eCommerce macht automatisierte Anwendungen im Shop-Management immer relevanter. Sehr hohes Potenzial wird der automatisierten Texterstellung in Online-Shops zugeschrieben. Der Beitrag stellt wichtige Merkmale der Produktbeschreibungen zusammen und vergleicht in einer empirischen Analyse automatisch und menschlich erstellte Produkttexte. Das Untersuchungsdesign berücksichtigt ein breites Spektrum an Erfolgskriterien für Produktbeschreibungen. Due to the increasing relevance of e‑commerce, automated applications for online shops are becoming more and more important. Particularly high potential is attributed to automated text creation. This article presents text characteristics that matter in product descriptions and compares automated and human-written texts in an empirical analysis. The research design takes into account a broad spectrum of success criteria for the specific text type.
  • Konferenzbeitrag
    On the State of German (Abstractive) Text Summarization
    (BTW 2023, 2023) Aumiller, Dennis; Fan, Jing; Gertz, Michael
    With recent advancements in the area of Natural Language processing, the focus is slowly shifting from a purely English-centric view towards more language-specific solutions, including German.Especially practical for businesses to analyze their growing amount of textual data are text summarization systems, which transform long input documents into compressed and more digestible summary texts.In this work, we assess the particular landscape of German abstractive text summarization and investigate the reasons why practically useful solutions for abstractive text summarization are still absent in industry. Our focus is two-fold, analyzing a) training resources, and b) publicly available summarization systems.We are able to show that popular existing datasets exhibit crucial flaws in their assumptions about the original sources, which frequently leads to detrimental effects on system generalization and evaluation biases. We confirm that for the most popular training dataset, MLSUM, over 50% of the training set is unsuitable for abstractive summarization purposes. Furthermore, available systems frequently fail to compare to simple baselines, and ignore more effective and efficient extractive summarization approaches. We attribute poor evaluation quality to a variety of different factors, which are investigated in more detail in this work:A lack of qualitative (and diverse) gold data considered for training, understudied (and untreated) positional biases in some of the existing datasets, and the lack of easily accessible and streamlined pre-processing strategies or analysis tools. We therefore provide a comprehensive assessment of available models on the cleaned versions of datasets, and find that this can lead to a reduction of more than 20 ROUGE-1 points during evaluation. As a cautious reminder for future work, we finally highlight the problems of solely relying on n-gram based scoring methods by presenting particularly problematic failure cases. Code for dataset filtering and reproducing results can be found online: https://github.com/anonymized-user/anonymized-repository
  • Textdokument
    Original oder Plagiat? Der schnelle Weg zur wissenschaftlichen Arbeit im Zeitalter künstlicher Intelligenz
    (INFORMATIK 2020, 2021) Weßels, Doris; Meyer, Eike
    Hochschulen müssen trotz der immer wieder artikulierten Ressourcen-und Kapazitätsprobleme zukunftsfähige Leitlinien und Praktiken für den Umgang mit studentischen Leistungen in Form schriftlicher Haus-und Abschlussarbeiten entwickeln – bis hin zur Entwicklung alternativer Konzepte als Bestandteil neuer Lernarchitekturen. Für die Hochschulleitungen und Lehrenden ergeben sich zwei Fragestellungen. Zum einen: Welchen Impact haben KI-basierte Werkzeuge des „Natural Language Generation bzw. Processing“ (NLG/NLP) für Prüfungsleistungen in Form schriftlicher Haus-und Abschlussarbeiten? Zum zweiten: Wie ist das „System Hochschule“ anzupassen, um seinem Bildungs-und Qualitätsanspruch im digitalen Zeitalter gerecht zu werden? In einem Online-Workshop im Rahmen der INFORMATIK2020 am 1.10.2020 wurden ausgewählte Werkzeuge in einem spielerisch anmutenden Team-Wettbewerb eingesetzt, um auf dieser Grundlage den schwierigen Grat zwischen Original und Plagiat bei der Erstellung wissenschaftlicher Arbeiten selbst zu erleben, das Verhalten als Lehrender und Forschender kritisch zu reflektieren und das Problembewusstsein für diese neue Herausforderung zu schärfen.