Refine
Year of publication
- 2015 (41) (remove)
Document Type
- Conference Proceeding (29)
- Article (3)
- Part of a Book (3)
- Part of Periodical (3)
- Contribution to a Periodical (2)
- Working Paper (1)
Is part of the Bibliography
- no (41) (remove)
Keywords
- Fachhochschule (2)
- Hochschule Ruhr West (2)
- Mülheim an der Ruhr (2)
- Zeitschrift (2)
- Data visualization (1)
- Framework (1)
- Image color analysis (1)
- Intercultural sharing (1)
- Knowledge sharing (1)
- Layout (1)
MeHRWert Ausgabe 6 März 2015
(2015)
In this work we report the first quasi-continuous in-situ photoluminescence study of growing InGaN LED structures inside an industrial-grade metal-organic vapor phase epitaxy (MOVPE) reactor at growth temperature. The photoluminescence spectra contain information about temperature, thickness and composition of the epitaxial layers. Furthermore, the in-situ spectra – even at an early stage of the growth of the active region – can be used to predict the photoluminescence emission wavelength of the structure at room temperature. In this study an accuracy of this predicted wavelength in the range of ± 1.3 nm (2σ) is demonstrated. This technique thus appears suitable for closed-loop control of the emission wavelength of InGaN LEDs already during growth.
Immer mehr ältere Menschen leben von ihren Angehörigen getrennt und können über Kommunikationsmedien wie Telefon und Skype nur eingeschränkt gemeinsame Erlebnisse erzeugen. In diesem Paper wird die technische Umsetzung eines Konzeptes vorgestellt, das es Familienmitgliedern ermöglicht über das Internet gemeinsam „Mensch-ärgere-dich-nicht“ zu spielen. Durch Videotelefonie und eine besondere Anordnung der Hardware werden die Spieler trotz räumlicher Trennung virtuell an einen Tisch gebracht und dadurch ein gemeinsames Erlebnis erzeugt. Die Clientanwendung wird dabei als plattformunabhängiger Webservice und die Videotelefonie mittels verschiedener Standards und Server realisiert.
4. Workshop Automotive HMI
(2015)
Benutzerschnittstellen im Fahrzeug stellen eine besondere Herausforderung in Konzeption und Entwicklung dar, steht doch eine sichere Bedienung in allen Fahrsituationen sowohl von Fahrerassistenzsystemen als auch von Komfort-und Unterhaltungsfunktionen im Vordergrund. Zugleich treffen durch zunehmende Vernetzung die langen Entwicklungszyklen von Kraftfahrzeugen auf die hochdynamische Welt von Mobiltelefonen und Internet. Ein-und Ausgabetechnologien gehören des Weiteren zu den zentralen Mitteln der Hersteller, die Wertigkeit der im Fahrzeug eingebauten Systeme hervorzuheben und sich gegenüber der Konkurrenz abzuheben. Dafür werden in diesem Workshop Konzepte und technische Lösungen von Designern, Entwicklern und Human Factors Experten aus Hochschulen, Forschungsinstituten und der Automobilindustrie vorgestellt und diskutiert.
In this demo paper we present a new visualization technique for dynamic networks. It displays the time slices of the dynamic network using two dimensional graph layouting algorithms and stacks these in the third dimension to show the development over time. The visualization ensures that the same node always has the same position in each time slice so that it is easy to follow its development. It also allows filtering data and influencing node appearance based on properties. Additionally we offer a two dimensional comparison view for two time slices which highlights changes in graph structure and (if available) in measures of nodes. The presented visualization technique is implemented using Web technology and is available in a Web-based analytics workbench. We demonstrate the benefits of these techniques by an analysis of a data set from a learning community.
Resource Usage in Online Courses: Analyzing Learner’s Active and Passive Participation Patterns
(2015)
The paper analyzes the experience with an open university course for a very heterogeneous target group in which MOOC-like materials and activities were used. The course was conducted in a specifically prepared and extended Moodle environment. The analysis involves questionnaires as well as performance data that reflect the resource access on the learning platform. A special focus is put on the participants’ acceptance and usage of student-generated versus teacher-provided learning content. Network analysis techniques have been used to identify "interest clusters" of students around certain resources.
This chapter describes our current research efforts related to the contextualization of learners in mobile learning activities. Substantial research in the field of mobile learning has explored aspects related to contextualized learning scenarios. However, new ways of interpretation and consideration of contextual information of mobile learners are necessary. This chapter provides an overview regarding the state of the art of innovative approaches for supporting contextualization in mobile learning. Additionally, we provide the description of the design and implementation of a flexible multi-dimensional vector space model to organize and process contextual data together with visualization tools for further analysis and interpretation. We also present a study with outcomes and insights on the usage of the contextualization support for mobile learners. To conlcude, we discuss the benefits of using contextualization models for learners in different use-cases. Moreover, a description is presented in order to illustrate how the proposed contextual model can easily be adapted and reused for different use-cases in mobile learning scenarios and potentially other mobile fields.
In the field of magnetic inductance tomography,
signal processing is a real challenge. This is due to the divergent
nature of magnetic fields. The sensitivity, i.e. the change in the
receiving signal by means of an electrically conductive sample
in a measuring volume depends strongly on the positioning
of the sample. Objects that are located near the transmitting
or receiving coils are very well locatable, where objects in
larger distance are hard to detect. In this paper an approach
is presented that improves the topology of the magnetic fields
in the ”magnetic induction tomography” (MIT) by changing
geometric constructions and current patterns of coils so far,
as to allow a sharper localization of objects within the space.
The aim is to level the distribution of the sensitivity in the
measuring volume, so that electrically conductive objects with
a larger distance between transmitting and receiving unit can
be detected with almost the same signal intensity as objects
close to the transmitting and receiving unit. The simulation tool
Comsolic is used for the geometric modeling making a finite
element analysis (FEA). The subsequent signal processing and
analysis of the simulation results are implemented in Matlabic .
Within this FEA the coil geometries and current patterns are
changed numerically, so that the minimum object size, that is
still detectable, is, compared to the known MIT, reduced and the
sensitivity of the system is improved. To validate the simulation in
Comsolic , first simulation results are compared with analytical
models and analyses.
Detection of influential actors in social media plays an important role for increasing the quality and efficiency of work and services in many fields such as education, marketing, etc. This work aims to introduce a new approach for the characterization of influential actors in online social media, such as Twitter. We present on a model of influence of an actor that is based on the attractiveness of the actor in terms of the number of other new actors with which he or she has established relations over time. We have used this concept and measure of influence to determine optimal seeds in a simulation of influence maximization using two empirically collected social networks for the underlying graphs.
Recently, rescue worker resources have not been sufficient to meet the regular response time during large-scale catastrophic events in every case. However, many volunteers supported official forces in different disaster situations, often self-organized through social media. In this paper, a system will be introduced which allows the coordination of trained volunteers by a professional control center with the objective of a more efficient distribution of human resources and technical equipment. Volunteers are contacted via app on their private smartphone. The design of this app is based on user requirements gathered in focus group discussions. The feedback of the potential users includes privacy aspects, low energy consumption, and mechanisms for long-term motivation and training. The authors present the results of the focus group analyses as well as the transfer to their app design concept.
In this paper we present an approach for People-to-People recommendations based on a Rich Context Model (RCM). We consider personal user information as contextual information used for our recommendations. The evaluation of our recommendation approach was performed on a social network of students. The obtained results do show a significant increase in performance while, at the same time, a slight increase in quality in comparison to a manual matching process. The proposed approach is flexible enough to handle different data types of contextual information and easy adaptable to other recommendation domains.
Why do barriers to the exchange of open knowledge resources change in public administrations? Experts in the public sector have been interviewed and outlined antecedents of change to certain barriers. The results are an initial step towards theorizing on barrier change and stepping beyond the current trend of categorizing difficulties to e-Learning and use of open knowledge resources. Categorizing only shows the range of potential challenges. Whether and how the barriers change, however, is seldom addressed in previous literature. The results presented in this study thus provide a new perspective on the phenomenon. Results are part of a longitudinal study about open e-Learning in the public sector across four European countries. They will provide fresh empirical input for discussions at the World Conference on E-Learning how to advance future research and practices in the domain
E-Learning and openness in education are receiving ever increasing attention in businesses as well as in academia. However, these practices have only to small extent been introduced in public administrations. The study addresses this gap by presenting a literature review on Open Educational Resources [OER] and E-Learning in the public sector. The main goal of the article is to identify challenges to open E-Learning in public administrations. Experiences will be conceptualized as barriers which need to be considered when introducing open E-Learning systems and programs in administrations. The main outcome is a systematic review of lessons learned, presented as a contextualized Barrier Framework which is suitable to analyze requirements when introducing E-Learning and OER in public administrations.
Die Entwicklung von vollautomatisierten Fahrzeugen wird in der gesellschaftlichen Diskussion immer präsenter. Wichtig für die Durchsetzung und verbreitete Nutzung dieser technischer Neuerungen ist jedoch vor allem die Akzeptanz der Bevölkerung – in diesem Fall nicht nur die der potenziellen KäuferInnen sondern auch die der übrigen Verkehrs-teilnehmenden. Vorgestellt wird eine explorative Online-Studie zur Akzeptanz von auto-nomen Fahren basierend auf quantitativen und qualitativen Daten einer Stichprobe von N = 89. Die Ergebnisse zeigen unter anderem eine geringe Vertrautheit mit dem Thema, ein vergleichsweise ausgeprägtes Vertrauen aber eine geringe Nutzungsabsicht.
Der Bedarf an feuerverzinkten Stahlbändern ist besonders in der Automobilindustrie sehr groß und es werden zugleich immer höhere Qualitäten gefordert. Hierbei bildet vor allem die Homogenität der Zinkschichtdicke ein entscheidendes Qualitätsmerkmal. Um das Stahlband ausreichend vor Umwelteinflüssen zu schützen, muss eine, vom Kunden spezifizierte, Mindestzinkschichtdicke aufgetragen werden. Beim hier angewandten Verzinkungsverfahren durchläuft das Band ein Zinkbad und anschließend wird das überschüssige Zink berührungs-los mittels sogenannten Abblasdüse so abgetragen, sodass eine möglichst homogene Zinkschicht erhalten bleibt. Hierzu ist es notwendig den Abstand zwischen Band und Airknife konstant zu halten. Störende Bandbewegungen führen zu inhomogene Zinkschichtdicken, welche die Qualität der Verzinkung vermindern. Diese Qualitätsverminderung und der erhöhte Zinkeinsatz soll durch geeignete Maßnahmen verringert werden. Bisher eingesetzte berührungslose Bandstabilisatoren können die Bandbewegung im Allge-meinen dämpfen, jedoch treten noch Betriebszustände auf, in denen eine inhomogene Zinkschicht sichtbar ist. Die Ursache dieser Inhomogenitäten liegt in anlagenbedingten dominanten Schwingungen des Bandes, deren Ursache zu klären ist. Im vorliegenden Beitrag wird ein Modell der Bandbewegung vorgestellt, das durch die theore-tische Modellbildung und experimentelle Identifikation erstellt worden ist. Das Modell beschreibt die Bewegung des Bandes bezüglich ausgewählter Freiheitsgrade und ermöglicht die Analyse der kritischen Betriebszustände. Darüber hinaus soll dieses Modell zur Stabilisierung des Bandes in einer modellgestützten Reglung verwendet werden.
As smart homes are being more and more popular, the needs of finding assisting systems which interface between users and home environments are growing. Furthermore, for people living in such homes, elderly and disabled people in particular and others in general, it is totally important to develop devices, which can support and aid them in their ordinary daily life. We focused in this work on sustaining privacy issues of the user during a real interaction with the surrounding home environment. A smart person-specific assistant system for services in home environment is proposed. The role of this system is the assisting of persons by controlling home activities and guiding the adaption of Smart-Home-Human interface towards the needs of the considered person. At the same time the system sustains privacy issues of it’s interaction partner. As a special case of medical assisting the system is so implemented, that it provides for elderly or disabled people person-specific medical assistance . The system has the ability of identifying its interaction partner using some biometric features. According to the recognized ID the system, first, adopts towards the needs of recognized person. Second the system represents person-specific list of medicines either visually or auditive. And third the system gives an alarm in the case of taking medicament either later or earlier as normal taking time.
We present a novel method to perform multi-class pattern classification with neural networks and test it on a challenging 3D hand gesture recognition problem. Our method consists of a standard one-against-all (OAA) classification, followed by another network layer classifying the resulting class scores, possibly augmented by the original raw input vector. This allows the network to disambiguate hard-to-separate classes as the distribution of class scores carries considerable information as well, and is in fact often used for assessing the confidence of a decision. We show that by this approach we are able to significantly boost our results, overall as well as for particular difficult cases, on the hard 10-class gesture classification task.
We present a novel approach of distributing matrix multiplications among GPU-equipped nodes in a cluster system. In this context we discuss the induced challenges and possible solutions. Additionally we state an algorithm which outperforms optimized GPU BLAS libraries for small matrices. Furthermore we provide a novel theoretical model for distributing algorithms within homogeneous computation systems with multiple hierarchies. In the context of this model we develop an algorithm which can find the optimal distribution parameters for each involved subalgorithm. We provide a detailed analysis of the algorithms space and time complexities and justify its use with a structured evaluation within a small GPU-equipped Beowulf cluster.
The influence of national culture on knowledge sharing has important implications for all organizations. However, the existing frameworks only cover a subset of relevant factors or limit the research of the framework to either organizational or national level. Hence, a more encompassing framework is needed. The question this articles answers is how does national culture influence knowledge sharing. Based on extensive literature review and interviews carried out in Finland and Japan, this article sets forth a foundation for a new framework. The framework details how national culture influences individual level and organizational level factors and technical tools. Additionally, the framework includes a new dimension, time-dimension, which is usually disregarded in knowledge sharing research. For researchers and practitioners, the derived framework provides key insight on relevant factors on knowledge sharing and national culture. Finally, future research directions are discussed.
This chapter describes our research efforts related to the design of mobile learning (m-learning) applications in cloud-computing (CC) environments. Many cloud-based services can be used/integrated in m-learning scenarios, hence, there is a rich source of applications that could easily be applied to design and deploy those within the context of cloud-based services. Here, we present two cloud-based approaches—a flexible framework for an easy generation and deployment of mobile learning applications for teachers, and a flexible contextualization service to support personalized learning environment for mobile learners. The framework provides a flexible approach that supports teachers in designing mobile applications and automatically deploys those in order to allow teachers to create their own m-learning activities supported by mobile devices. The contextualization service is proposed to improve the content delivery of learning objects (LOs). This service allows adapting the learning content and the mobile user interface (UI) to the current context of the user. Together, this leads to a powerful and flexible framework for the provisioning of potentially ad hoc mobile learning scenarios. We provide a description of the design and implementation of two proposed cloud-based approaches together with scenario examples. Furthermore, we discuss the benefits of using flexible and contextualized cloud applications in mobile learning scenarios. Hereby, we contribute to this growing field of research by exploring new ways for designing and using flexible and contextualized cloud-based applications that support m-learning.
Bei Großschadensereignissen kann es durch die Vielzahl der Alarme dazu kommen, dass die verfügbaren Rettungskräfte nicht mehr ausreichen, um die anfallenden Aufgaben zu bewältigen oder Hilfsfristen einzuhalten. Die vorliegende Arbeit beschreibt einen Ansatz, sich zusätzlicher Hilfe aus der Bevölkerung zu bedienen, die über einen Disponenten aus der vorhandenen Leitstelle koordiniert wird. Dabei stehen nicht spontan organisierte Helfer im Vordergrund, sondern Personen, die sich vorab mit einem klaren Fertigkeitsprofil und ggf. auch Ausstattung im System registriert haben. Besondere Anforderungen entstehen bei den Disponenten der Leitstelle, deren Mehrbelastung durch das neue System gering zu halten ist, als auch bei den freiwilligen Helfern, die über eine App auf dem Mobiltelefon alarmiert werden und auch darüber die Kommunikation führen sollen. Die Anforderungen beeinflussen sowohl die System-Infrastruktur als auch die Benutzerschnittstelle.
Mensch-Maschine-Interaktion in sicherheitskritischen Systemen ist ein für die Informatik und die jeweiligen Anwendungsdomänen in der Bedeutung weiter zunehmendes Thema. Dieser Workshop der GI-Fachgruppe „Mensch-Maschine-Interaktion in sicherheitskritischen Systemen" innerhalb des Fachbereichs Mensch-Computer-Interaktion soll aktuelle Entwicklungen und Fragestellungen offenlegen und neue Impulse für das Forschungsgebiet geben.
Social networking sites (SNSs) are an integral part of our daily life. With the evermore increasing appearance of SNSs, their users spend considerable time producing of different forms everyday (e.g. text, videos, photos, links, etc.) or browsing the varieties of contents in different SNSs. In this paper, we propose an architectural perspective on a framework that provides a unified environment through which users can produce and browse different contents of SNSs from one place.
With the spread of mobile devices among both, men and women, app-based games also become more popular. While traditionally, digital games are more famous among men, women seem to spend more time and money on mobile gaming. There are a lot of open questions with regard to women and gaming in general; research on gender differences in app-based mobile gaming is almost nonexistent. Taking an exploratory perspective, our study investigates gender differences in general usage patterns, attachment towards the game and motivational differences for choosing to play the famous QuizClash app. Also, we identify differences in reported and actual performance in specific categories and capture anticipation of success as well as likeliness of choosing specific knowledge categories depending on the opponents’ performance profile.
We present a novel hierarchical approach to multi-class classification which is generic in that it can be applied to different classification models (e.g., support vector machines, perceptrons), and makes no explicit assumptions about the probabilistic structure of the problem as it is usually done in multi-class classification. By adding a cascade of additional classifiers, each of which receives the previous classifier's output in addition to regular input data, the approach harnesses unused information that manifests itself in the form of, e.g., correlations between predicted classes. Using multilayer perceptrons as a classification model, we demonstrate the validity of this approach by testing it on a complex ten-class 3D gesture recognition task.
This paper discusses the efforts carried out related to the design and development of a web-based framework that allows designing, deploying and executing mobile data collecting applications. Furthermore, it also allows analyzing and presenting the data that is generated during the mentioned process. The fact that the framework is completely web-based provides a platform independent execution of the mobile application on any mobile device with a web-browser. As a result that the whole life-cycle of creating, executing and discussing a mobile learning activity is implemented in pure web-based manner separates this work from similar efforts. In the course of this work, the current state of development of two of the components, the authoring tool and the mobile application is presented. This framework was introduced to teachers in an activity to follow up an initial study. On the basis of a workshop with teachers, we performed an explorative study regarding the technology acceptance and usability of two components of the proposed framework. The results are discussed and analyzed in this paper.
Global software development changes the requirements in terms of soft competency and increases the complexity of social interaction by including intercultural aspects. While soft competency is often seen as crucial for the success of global software development projects, the concrete competence requirements remain unknown. Internationalization competency represents one of the first attempts to structure and describe the soft competence requirements for global software developers. Based on the diversity of tasks, competence requirements will differ among the various phases of software development. By conducting a survey on the importance of internationalization competences for the different phases of global software development, we identified differences in terms of competence importance and requirements in the phases. “Adaptability” (of one's working style) and “Cultural Awareness” were the main differences. “Cultural Awareness” distinguishes requirements engineering and software design from testing and implementation while “Adaptability” distinguishes implementation and software design from requirements engineering and testing.
The continuous evolution of learning technologies combined with the changes within ubiquitous learning environments in which they operate result in dynamic and complex requirements that are challenging to meet. The fact that these systems evolve over time makes it difficult to adapt to the constant changing requirements. Existing approaches in the field of Technology Enhanced Learning (TEL) are generally not addressing those issues and they fail to adapt to those dynamic situations. In this chapter, we investigate the notion of an adaptive and adaptable architecture as a possible solution to address these challenges. We conduct a literature survey upon the state of the art and state of practice in this area. The outcomes of those efforts result in an initial model of a Domain-specific architecture to tackle the issues of adaptability and adaptiveness. To illustrate these ideas, we provide a number of scenarios where this architecture can be applied or is already applied. Our proposed approach serves as a foundation for addressing future ubiquitous learning applications where new technologies constantly emerge and new requirements evolve.
Utilizing biometrie traits for privacy- and security-applications is receiving an increasing attention. Applications such as personal identification, access control, forensics appli-cations, e-banking, e-government, e-health and recently person-alized human-smart-home and human-robot interaction present some examples. In order to offer person-specific services for/of specific person a pre-identifying step should be done in the run-up. Using biometric in such application is encountered by diverse challenges. First, using one trait and excluding the others depends on the application aimed to. Some applications demand directly touch to biometric sensors, while others don't. Second challenge is the reliability of used biometric arrangement. Civilized application demands lower reliability comparing to the forensics ones. And third, for biometric system could only one trait be used (uni-modal systems) or multiple traits (Bi- or Multi-modal systems). The latter is applied, when systems with a relative high reliability are expected. The main aim of this paper is providing a comprehensive view about biometric and its application. The above mentioned challenges will be analyzed deeply. The suitability of each biometric sensor according to the aimed application will be deeply discussed. Detailed com-parison between uni-modal and Multi-modal biometric system will present which system where to be utilized. Privacy and security issues of biometric systems will be discussed too. Three scenarios of biometric application in home-environment, human-robot-interaction and e-health will be presented.
A light-weight real-time ap- plicable hand gesture recognition system for automotive applications
(2015)
We present a novel approach for improved hand-gesture recognition by a single time-of-flight(ToF) sensor in an automotive environment. As the sensor's lateral resolution is comparatively low, we employ a learning approach comprising multiple processing steps, including PCA-based cropping, the computation of robust point cloud descriptors and training of a Multilayer perceptron (MLP) on a large database of samples. A sophisticated temporal fusion technique boosts the overall robustness of recognition by taking into account data coming from previous classification steps. Overall results are very satisfactory when evaluated on a large benchmark set of ten different hand poses, especially when it comes to generalization on previously unknown persons.
We present a system for efficient dynamic hand gesture recognition based on a single time-of-flight sensor. As opposed to other approaches, we simply rely on depth data to interpret user movement with the hand in mid-air. We set up a large database to train multilayer perceptrons (MLPs) which are subsequently used for classification of static hand poses that define the targeted dynamic gestures. In order to remain robust against noise and to balance the low sensor resolution, PCA is used for data cropping and highly descriptive features, obtainable in real-time, are presented. Our simple yet efficient definition of a dynamic hand gesture shows how strong results are achievable in an automotive environment allowing for interesting and sophisticated applications to be realized.
This contribution demonstrates the efficient embedding of a single depth-camera into the automotive environment making mid-air gesture interaction for mobile applications viable in such a scenario. In this setting a new human-machine interface is implemented to give an idea of future improvements in automation processes in industrial applications. Our system is based on a data-driven approach by learning hand poses as well as gestures from a large database in order to apply them on mobile devices. We register any movement in a nearby driver area and crop data efficiently with the means of PCA transforming it into so-called feature vectors which present the input for our multi-layer perceptrons (MLPs). After MLP classification, the interpretation of user input is sent via WiFi to a tablet PC mounted into the car interior visualizing an infotainment system which the user is able to interact with. We demonstrate that by this setup hand gestures as well as hand poses are easily and efficiently interpretable insofar as that they become an intuitive and supplementary means of interaction for automotive HMI in mobile scenarios realizable in real-time.
Human computer interaction in security and time-critical systems is an interdisciplinary challenge at the seams of human factors, engineering, information systems and computer science. Application fields include control systems, critical infrastructures, vehicle and traffic management, production technology, business continuity management, medical technology, crisis management and civil protection. Nowadays in many areas mobile and ubiquitous computing as well as social media and collaborative technologies also plays an important role. The specific challenges require the discussion and development of new methods and approaches in order to design information systems. These are going to be addressed in this special issue with a particular focus on technologies for citizen and volunteers in emergencies.
Object detection systems which operate on large data streams require an efficient scaling with available computation power. We analyze how the use of tile-images can increase the efficiency (i.e. execution speed) of distributed HOG-based object detectors. Furthermore we discuss the challenges of using our developed algorithms in practical large scale scenarios. We show with a structured evaluation that our approach can provide a speed-up of 30-180 % for existing architectures. Due to the its generic formulation it can be applied to a wide range of HOG-based (or similar) algorithms. In this context we also study the effects of applying our method to an existing detector and discuss a scalable strategy for distributing the computation among nodes in a cluster system.
Forschung an Hochschulen
(2015)
In diesem Aufsatz soll die Forschung an Fachhochschulen beispielhaft aus dem Blickwinkel des Instituts Informatik der in 2009 gegründeten Hochschule Ruhr West betrachtet werden. Am Institut Informatik ist es das Ziel Lehre und Forschung geeignet zu verknüpfen, um Studierenden, wissenschaftlichen Mitarbeiterinnen und Mitarbeitern und auch Lehrenden ein attraktives Angebot in Forschung und Lehre im Bereich der Informatik zu liefern. Dabei bilden neben der Durchführung interessanter Lehrveranstaltungen, welche durch aktuelle Forschungsfragestellungen angereichert werden, das kooperative Bearbeiten von gesellschaftlich relevanten und zukunftsweisenden Forschungsaufgaben, die Teilnahme an Forschungsverbünden, bilaterale Forschungsaktivitäten mit Partnern aus der Wirtschaft und das Einwerben von externen Mitteln, die Basis der Arbeit am Institut.
We present a novel approach of distributing small-to mid-scale neural networks onto modern parallel architectures. In this context we discuss the induced challenges and possible solutions. We provide a detailed theoretical analysis with respect to space and time complexities and reinforce our computation model with evaluations which show a performance gain over state of the art approaches.