Fraunhofer-Institut für Optronik, Systemtechnik und Bildauswertung IOSB

Filters

3 1
Reset filters

Settings
Now showing 1 - 4 of 4
  • Publication
    Best-of-Breed: Service-Oriented Integration of Artificial Intelligence in Interoperable Educational Ecosystems
    ( 2023)
    Krauß, Christopher orcid-logo
    ;
    Streicher, Alexander
    ;
    Poxleitner, Eva
    ;
    Altun, Daniela
    ;
    Müller, Joanna
    ;
    An, Truong-Sinh
    ;
    Müller, Christoph
    Artificial Intelligence (AI) offers great potential for optimizing learning processes, teaching methods, learning content, or organizational procedures. However, the success of AI components in educational environments is by no means guaranteed and depends on several conditions in their respective learning settings. In this article, we analyze requirements that are often addressed prior to introducing AI features. We address organizational, methodological, didactical, content-related, and technical challenges. The research question of this work is how AI features can best be incorporated into modern educational system landscapes to create sustainable system architectures that are accepted and perceived as added value by users. Thereby, the article discusses two approaches to software architecture: Best-of-Suite (for monolithic architectures) and Best-of-Breed (for service-oriented architectures). Monolithic systems offer a wide range of functions, can be offered by a single provider but can become difficult to manage and create dependencies. Specialized and service-oriented systems, in turn, consist of modular functions handled by specialized services, are more flexible and scalable, and can be integrated with a wide range of tools and services, but require more effort to set up and manage. We explain why the Best-of-Breed strategy is a sensible approach to the use of AI components, how this can be implemented sustainably with the help of a middleware component, and we report on the user experiences from a field test. While in this work we evaluate the implemented system with a cybersecurity training as an on-the-job course, the middleware has been successfully used in other educational contexts, as well.
  • Publication
    Lessons learned from creating, implementing and evaluating assisted e-learning incorporating adaptivity, recommendations and learning analytics
    ( 2022)
    Altun, Daniela
    ;
    Krauß, Christopher orcid-logo
    ;
    Streicher, Alexander
    ;
    Müller, Christoph
    ;
    Atorf, Daniel
    ;
    Rerhaye, Lisa
    ;
    Kunde, Dietmar
    Applications of adaptive e-learning, recommender systems and learning analytics are typically presented individually, however, their combination poses several challenging requirements ranging from organizational to technical issues. This article presents a technical study from a holistic application of a variety of e-learning assistance technologies, including recommender systems, chatbots, adaptivity, and learning analytics. At its core we operationalize interoperability standards such as the Experience API (xAPI) and Learning Tools Interoperability (LTI), and control the data flow via a standard-encapsulating middleware approach. We report on the challenges regarding organization, methodology, content, didactics, and technology. A systematic evaluation with the target group discusses the users’ expectations with the measured interactions.
  • Publication
    Generative Machine Learning for Resource-Aware 5G and IoT Systems
    ( 2021)
    Piatkowski, Nico
    ;
    Mueller-Roemer, Johannes Sebastian orcid-logo
    ;
    Hasse, Peter
    ;
    Bachorek, Adam
    ;
    Werner, Tim
    ;
    Birnstill, Pascal
    ;
    Morgenstern, Andreas
    ;
    Stobbe, Lutz
    Extrapolations predict that the sheer number of Internet-of-Things (IoT) devices will exceed 40 billion in the next five years. Hand-crafting specialized energy models and monitoring sub-systems for each type of device is error prone, costly, and sometimes infeasible. In order to detect abnormal or faulty behavior as well as inefficient resource usage autonomously, it is of tremendous importance to endow upcoming IoT and 5G devices with sufficient intelligence to deduce an energy model from their own resource usage data. Such models can in-turn be applied to predict upcoming resource consumption and to detect system behavior that deviates from normal states. To this end, we investigate a special class of undirected probabilistic graphical model, the so-called integer Markov random fields (IntMRF). On the one hand, this model learns a full generative probability distribution over all possible states of the system-allowing us to predict system states and to measure the probability of observed states. On the other hand, IntMRFs are themselves designed to consume as less resources as possible-e.g., faithful modelling of systems with an exponentially large number of states, by using only 8-bit unsigned integer arithmetic and less than 16KB memory. We explain how IntMRFs can be applied to model the resource consumption and the system behavior of an IoT device and a 5G core network component, both under various workloads. Our results suggest, that the machine learning model can represent important characteristics of our two test systems and deliver reasonable predictions of the power consumption.
  • Publication
    OpenIoT: Open source Internet-of-Things in the cloud
    ( 2015)
    Soldatos, J.
    ;
    Kefalakis, N.
    ;
    Hauswirth, M.
    ;
    Serrano, M.
    ;
    Calbimonte, J.-P.
    ;
    Riahi, M.
    ;
    Aberer, K.
    ;
    Jayaraman, P.P.
    ;
    Zaslavsky, A.
    ;
    Zarko, I.
    ;
    Skorin-Kapov, L.
    ;
    Herzog, Reinhard
    Despite the proliferation of Internet-of-Things (IoT) platforms for building and deploying IoT applications in the cloud, there is still no easy way to integrate heterogeneous geographically and administratively dispersed sensors and IoT services in a semantically interoperable fashion. In this paper we provide an overview of the OpenIoT project, which has developed and provided a first-of-kind open source IoT platform enabling the semantic interoperability of IoT services in the cloud. At the heart of OpenIoT lies the W3C Semantic Sensor Networks (SSN) ontology, which provides a common standards-based model for representing physical and virtual sensors. OpenIoT includes also sensor middleware that eases the collection of data from virtually any sensor, while at the same time ensuring their proper semantic annotation. Furthermore, it offers a wide range of visual tools that enable the development and deployment of IoT applications with almost zero programming. Another key feature of OpenIoT is its ability to handle mobile sensors, thereby enabling the emerging wave of mobile crowd sensing applications. OpenIoT is currently supported by an active community of IoT researchers, while being extensively used for the development of IoT applications in areas where semantic interoperability is a major concern.