1.
Podlesny, N.J., Kayem, A.V., Meinel, C.: Towards Identifying De-anonymisation Risks in Distributed Health Data Silos. International Conference on Database and Expert Systems Applications. pp. 33–43. Springer (2019).
2.
von Schmieden, K., Staubitz, T., Mayer, L., Meinel, C.: Skill Confidence Ratings in a MOOC: Examining the Link Between Skill Confidence and Learner Development. Proceedings of the 11th International Conference on Computer Supported Education. pp. 533–40 (2019).
3.
Najafi, P., Mühle, A., Pünter, W., Cheng, F., Meinel, C.: MalRank: A Measure of Maliciousness in SIEM-based Knowledge Graphs. Proceedings of the 35th Annual Computer Security Applications Conference. pp. 417–429. ACM (2019).
In this paper, we formulate threat detection in SIEM environments as a large-scale graph inference problem. We introduce a SIEM- based knowledge graph which models global associations among entities observed in proxy and DNS logs, enriched with related open-source intelligence (OSINT) and cyber threat intelligence (CTI). Next, we propose MalRank, a graph-based inference algorithm designed to infer a node maliciousness score based on its associations to other entities presented in the knowledge graph, e.g., shared IP ranges or name servers. After a series of experiments on real-world data captured from a global enterprise’s SIEM (spanning over 3TB of disk space), we show that MalRank maintains a high detection rate (AUC = 96%) outperforming its predecessor, Belief Propagation, both in terms of accuracy and efficiency. Furthermore, we show that this approach is effective in identifying previously unknown malicious entities such as malicious domain names and IP addresses. The system proposed in this research can be implemented in conjunction with an organization’s SIEM, providing a maliciousness score for all observed entities, hence aiding SOC investigations.
4.
Podlesny, N.J., Kayem, A.V., Meinel, C.: Identifying Data Exposure Across Distributed High-Dimensional Health Data Silos through Bayesian Networks Optimised by Multigrid and Manifold. 2019 IEEE Intl Conf on Dependable, Autonomic and Secure Computing, Intl Conf on Pervasive Intelligence and Computing, Intl Conf on Cloud and Big Data Computing, Intl Conf on Cyber Science and Technology Congress (DASC/PiCom/CBDCom/CyberSciTech). pp. 556–563. IEEE (2019).
9.
von Schmieden, K., Meinel, C.: Utilizing Warm-Up Games in MOOC Discussion Forums. EMOOCs-WiP. pp. 218–223 (2019).
10.
von Schmieden, K., Mayer, L., Meinel, C.: Learner Response to Brainstorming Techniques in a Design Thinking MOOC. Cumulus Conference Proceedings. pp. 443–455. Cumulus Design Conference (2019).
11.
Rohloff, T., Renz, J., Suarez, G.N., Meinel, C.: A Ubiquitous Learning Analytics Architecture for a Service-Oriented MOOC Platform. In: Calise, M., Delgado Kloos, C., Reich, J., Ruiperez-Valiente, J.A., and Wirsing, M. (eds.) Digital Education: At the MOOC Crossroads Where the Interests of Academia and Business Converge (EMOOCs 2019). pp. 162–171. Springer International Publishing (2019).
As Massive Open Online Courses (MOOCs) generate a huge amount of learning activity data through its thousands of users, great potential is provided to use this data to understand and optimize the learning experience and outcome, which is the goal of Learning Analytics. But first, the data needs to be collected, processed, analyzed and reported in order to gain actionable insights. Technical concepts and implementations are rarely accessible and therefore this work presents an architecture how Learning Analytics can be implemented in a service-oriented MOOC platform. To achieve that, a service based on extensible schema-agnostic processing pipelines is introduced for the HPI MOOC platform. The approach was evaluated regarding its scalability, extensibility, and versatility with real-world use cases. Also, data privacy was taken into account. Based on five years of running the service in production on several platform deployments, six design recommendations are presented which can be utilized as best practices for platform vendors and researchers when implementing Learning Analytics in MOOCs.
12.
Grüner, A., Mühle, A., Meinel, C.: An Integration Architecture to Enable Service Providers for Self-sovereign Identity. Proceedings of the 18th. International Symposium on Network Computing and Applications. IEEE, Boston, MA (2019).
The self-sovereign identity management model emerged with the rise of blockchain technology. This paradigm focuses on user-centricity and strives to place the user in full control of the digital identity. Numerous implementations embrace the self-sovereign identity concept, leading to a fragmented landscape of solutions. At the same time, traditional identity and access management protocols are largely disregarded and facilities to issue verifiable claims as attributes are not available. Therefore, service providers barely adopt these solutions. We propose a component-based architecture for integrating selfsovereign identity solutions into web applications to foster their adoption by service providers. Furthermore, we outline a sample implementation as a gateway that enables uPort and Jolocom for authentication, via the OpenID Connect protocol, as well as the retrieval of email address attestations for these solutions.
13.
Torkura, K. .A, Sukmana, M.I.H., Cheng, F., Meinel, C.: Security Chaos Engineering for Cloud Services. The Proceedings of 18th IEEE International Symposium on Network Computing and Applications (NCA 2019). IEEE (2019).
14.
Podlesny, N.J., Kayem, A.V., Meinel, C., Jungmann, S.: How Data Anonymisation Techniques influence Disease Triage in Digital Health: A Study on Base Rate Neglect. Proceedings of the 9th International Conference on Digital Public Health. pp. 55–62. ACM (2019).
15.
Rohloff, T., Sauer, D., Meinel, C.: Student Perception of a Learner Dashboard in MOOCs to Encourage Self-Regulated Learning. IEEE International Conference on Engineering, Technology and Education (TALE 2019). IEEE (2019).
In online learning environments like Massive Open Online Courses (MOOCs), where teachers cannot provide individual support and guidance for thousands of students, self-regulated learning (SRL) is a critical metacognitive skillset for students’ achievement. However, not every student intuitively self-regulates its learning and therefore technical solutions can help to apply SRL strategies. Learner dashboards with visualizations about the learner’s progress and behavior are able to create awareness, encourage self-reflection, and perhaps motivate students to plan and adjust their learning behavior to achieve their learning objectives. Hence, such Learning Analytics tools can support the SRL strategies self-evaluation and strategic planning. To examine this potential, a learner dashboard was integrated into the HPI MOOC platform. This work presents the design process, the concept, and an evaluation of the first dashboard iteration. The perceived usefulness and usability are investigated, and in addition, the question will be considered whether the dashboard encourages students to apply self-regulated learning. The positive results pave the way for future research and a next iteration of the learner dashboard.
16.
Bothe, M., Rohloff, T., Meinel, C.: A Quantitative Study on the Effects of Learning with Mobile Devices in MOOCs. 2019 IEEE International Conference on Engineering, Technology and Education (TALE). pp. 1–7 (2019).
Massive Open Online Course (MOOC) platforms were initially designed for a desktop learning experience delivered via the Internet. With the increasing acceptance of mobile devices, learners started accessing the MOOC platforms through the browser application on their smartphones and tablets. However, native mobile applications offer better system integration and enhance the learning experience. As the concept of mobile-assisted seamless learning emphasizes the ubiquitous access to learning material, the relevance of mobile devices in the learning process will increase further. This paper investigates the different learning behaviors when using mobile devices on the HPI MOOC platform. For this, influencing aspects, that can not always be controlled by the learner, are examined for native applications and mobile websites-such as the size of the screen and the current network state of the mobile device. The results of a quantitative study show highly significant differences between the usage of native applications, mobile websites, and the overall average of the HPI MOOC platform. It was proven that the size of the screen has a large practical effect when using native applications. Furthermore, course items and videos are more often consumed when the device is connected to a WiFi network. This study creates the basis for future research to improve the support of mobile-assisted seamless learning methods for MOOCs.
17.
Tietz, C., Klieme, E., Behrendt, L., Böning, P., Marschke, L., Meinel, C.: Verification of Keyboard Acoustics Authentication on Laptops and Smartphones Using WebRTC. 2019 3rd Cyber Security in Networking Conference (CSNet). pp. 130–137 (2019).
18.
Bartz, C., Yang, H., Bethge, J., Meinel, C.: LoANs: Weakly Supervised Object Detection with Localizer Assessor Networks. In: Carneiro, G. and You, S. (eds.) Computer Vision -- ACCV 2018 Workshops. pp. 341–356. Springer International Publishing, Perth, Australia (2019).
Recently, deep neural networks have achieved remarkable performance on the task of object detection and recognition. The reason for this success is mainly grounded in the availability of large scale, fully annotated datasets, but the creation of such a dataset is a complicated and costly task. In this paper, we propose a novel method for weakly supervised object detection that simplifies the process of gathering data for training an object detector. We train an ensemble of two models that work together in a student-teacher fashion. Our student (localizer) is a model that learns to localize an object, the teacher (assessor) assesses the quality of the localization and provides feedback to the student. The student uses this feedback to learn how to localize objects and is thus entirely supervised by the teacher, as we are using no labels for training the localizer. In our experiments, we show that our model is very robust to noise and reaches competitive performance compared to a state-of-the-art fully supervised approach. We also show the simplicity of creating a new dataset, based on a few videos (e.g. downloaded from YouTube) and artificially generated data.
19.
Rezaei, M., Yang, H., Harmuth, K., Meinel, C.: Conditional generative adversarial refinement networks for unbalanced medical image semantic segmentation. 2019 IEEE winter conference on applications of computer vision (WACV). pp. 1836–1845. IEEE (2019).
20.
Staubitz, T., Meinel, C.: Graded Team Assignments in MOOCs: Effects of Team Composition and Further Factors on Team Dropout Rates and Performance. Proceedings of the Sixth (2019) ACM Conference on Learning @ Scale. pp. 5:1–5:10. ACM, Chicago, IL, USA (2019).
The ability to work in teams is an important skill in today's work environments. In MOOCs, however, team work, team tasks, and graded team-based assignments play only a marginal role. To close this gap, we have been exploring ways to integrate graded team-based assignments in MOOCs. Some goals of our work are to determine simple criteria to match teams in a volatile environment and to enable a frictionless online collaboration for the participants within our MOOC platform. The high dropout rates in MOOCs pose particular challenges for team work in this context. By now, we have conducted 15 MOOCs containing graded team-based assignments in a variety of topics. The paper at hand presents a study that aims to establish a solid understanding of the participants in the team tasks. Furthermore, we attempt to determine which team compositions are particularly successful. Finally, we examine how several modifications to our platform's collaborative toolset have affected the dropout rates and performance of the teams.
21.
von Schmieden, K., Mayer, L., Taheri, M., Meinel, C.: Iterative course design in MOOCs: Evaluating a protoMOOC. Proceedings of the Design Society: International Conference on Engineering Design. pp. 539–548. Cambridge University Press (2019).
User-generated content on social media platforms is a rich source of latent information about individual variables. Crawling and analyzing this content provides a new approach for enterprises to personalize services and put forward product recommendations. In the past few years, brands made a gradual appearance on social media platforms for advertisement, customers support and public relation purposes and by now it became a necessity throughout all branches. This online identity can be represented as a brand personality that reflects how a brand is perceived by its customers. We exploited recent research in text analysis and personality detection to build an automatic brand personality prediction model on top of the (Five-Factor Model) and (Linguistic Inquiry and Word Count) features extracted from publicly available benchmarks. The proposed model reported significant accuracy in predicting specific personality traits form brands. For evaluating our prediction results on actual brands, we crawled the Facebook API for 100k posts from the most valuable brands’ pages in the USA and we visualize exemplars of comparison results and present suggestions for future directions.
23.
Grüner, A., Mühle, A., Gayvoronskaya, T., Meinel, C.: A Comparative Analysis of Trust Requirements in Decentralized Identity Management. Proceedings of the 33rd. International Conference on Advanced Information Networking and Applications. Springer, Matsue, Japan (2019).
Identity management is a fundamental component in securing online services. Isolated and centralized identity models have been applied within organizations. Moreover, identity federations connect digital identities across trust domain boundaries. These traditional models have been thoroughly studied with regard to trust requirements. The recently emerging blockchain technology enables a novel decentralized identity management model that targets user-centricity and eliminates the identity provider as a trusted third party. The result is a substantially different set of entities with mutual trust requirements. In this paper, we analyze decentralized identity management based on blockchain through defining topology patterns. These patterns depict schematically the decentralized setting and its main actors. We study trust requirements for the devised patterns and, finally, compare the result to traditional models. Our contribution enables a clear view of differences in trust requirements within the various models.
24.
Podlesny, N.J., Kayem, A.V., Meinel, C.: Attribute Compartmentation and Greedy UCC Discovery for High-Dimensional Data Anonymization. Proceedings of the Ninth ACM Conference on Data and Application Security and Privacy. pp. 109–119. ACM (2019).
25.
von Schmieden, K., Staubitz, T., Mayer, L., Meinel, C.: Skill Confidence Ratings in a MOOC: Examining the Link between Skill Confidence and Learner Development. CSEDU (2019).
27.
Podlesny, N.J.: High-dimensional data anonymization for in-memory applications, (2019).
28.
Bothe, M., Meinel, C.: Applied Mobile-Assisted Seamless Learning Techniques in MOOCs. In: Calise, M., Delgado Kloos, C., Reich, J., Ruiperez-Valiente, J.A., and Wirsing, M. (eds.) Digital Education: At the MOOC Crossroads Where the Interests of Academia and Business Converge. pp. 21–30. Springer International Publishing, Cham (2019).
As Massive Open Online Courses (MOOCs) are nowadays used in an increasingly ubiquitous manner, the learning process gets disrupted every time learners change context. Mobile-Assisted Seamless Learning (MSL) techniques have been identified to reduce unwanted overhead for learners and streamline their learning process. However, technical implementations vary across the industry. This paper examines existing MSL research and applied techniques in the context of MOOCs. Therefore, we discussed related MSL research topics. Afterward, eleven characteristic MSL features were selected and compared their implementations across five major MOOC platforms. While web applications provide a bigger feature set, mobile clients offer advanced offline capabilities. Based on the findings, a concept outlines how MSL features can enhance the learning experience on MOOC platforms while considering the technical feasibility.
29.
Torkura, K. .A, Sukmana, M.I.H., Cheng, F., Meinel, C.: SlingShot: Automated Threat Detection and Incident Response in Multi-Cloud Storage Systems. The Proceedings of 18th IEEE International Symposium on Network Computing and Applications (NCA 2019). IEEE (2019).
30.
Rohloff, T., Oldag, S., Renz, J., Meinel, C.: Utilizing Web Analytics in the Context of Learning Analytics for Large-Scale Online Learning. IEEE Global Engineering Education Conference (EDUCON 2019). pp. 296–305. IEEE (2019).
Today, Web Analytics (WA) is commonly used to obtain key information about users and their behavior on websites. Besides, with the rise of online learning, Learning Analytics (LA) emerged as a separate research field for collecting and analyzing learners’ interactions on online learning platforms. Although the foundation of both methods is similar, WA has not been profoundly used for LA purposes. However, especially large-scale online learning environments may benefit from WA as it is more sophisticated and well-established in comparison to LA. Therefore, this paper aims to examine to what extent WA can be utilized in this context, without compromising the learners’ data privacy. For this purpose, Google Analytics was integrated into the Massive Open Online Course platform of the Hasso Plattner Institute as a proof of concept. It was tested with two deployments of the platform: openHPI and openSAP, where thousands of learners gain academic and industry knowledge about engineering education. Besides capturing behavioral data, the platforms’ existing LA dashboards were extended by WA metrics. The evaluation of the integration showed that WA covers a large part of the relevant metrics and is particularly suitable for obtaining an overview of the platform’s global activity, but reaches its limitations when it comes to learner-specific metrics.
31.
Grüner, A., Mühle, A., Meinel, C.: Using Probabilistic Attribute Aggregation for Increasing Trust in Attribute Assurance. Proceedings of the 2019 IEEE Symposium Series on Computational Intelligence in Cyber Security. IEEE, Xiamen, China (2019).
Identity management is an essential cornerstone of securing online services. Service provisioning relies on correct and valid attributes of a digital identity. Therefore, the identity provider is a trusted third party with a specific trust requirement towards a verified attribute supply. This trust demand implies a significant dependency on users and service providers. We propose a novel attribute aggregation method to reduce the reliance on one identity provider. Trust in an attribute is modelled as a combined assurance of several identity providers based on probability distributions. We formally describe the proposed aggregation model. The resulting trust model is implemented in a gateway that is used for authentication with self-sovereign identity solutions. Thereby, we devise a service provider specific web of trust that constitutes an intermediate approach bridging a global hierarchical model and a locally decentralized peer to peer scheme.
32.
Rohloff, T., Sauer, D., Meinel, C.: On the Acceptance and Usefulness of Personalized Learning Objectives in MOOCs. Proceedings of the Sixth ACM Conference on Learning at Scale (L@S 2019). pp. 4:1–4:10. Association for Computing Machinery (2019).
With Massive Open Online Courses (MOOCs) the number of people having access to higher education increased rapidly. The intentions to enroll for a specific course vary significantly and depend on one's professional or personal learning needs and interests. All learners have in common that they pursue their individual learning objectives. However, predominant MOOC platforms follow a one-size-fits-all approach and primarily aim for completion with certification. Specifically, technical support for goal-oriented and self-regulated learning to date is very limited in this context although both learning strategies are proven to be key factors for students' achievement in large-scale online learning environments. In this first investigation, a concept for the application and technical integration of personalized learning objectives in a MOOC platform is realized and assessed. It is evaluated with a mixed-method approach. First, the learners' acceptance is examined with a multivariate A/B test in two courses. Second, a survey was conducted to gather further feedback about the perceived usefulness, next to the acceptance. The results show a positive perception by the learners, which paves the way for future research.
33.
Rezaei, M., Yang, H., Meinel, C.: Learning Imbalanced Semantic Segmentation through Cross-Domain Relations of Multi-Agent Generative Adversarial Networks. SPIE Medical Imaging - Computer Aided Diagnosis (SPIE 2019) (2019).
34.
Seidel, F., Krentz, K.-F., Meinel, C.: Deep En-Route Filtering of Constrained Application Protocol (CoAP) Messages on 6LoWPAN Border Routers. Proceedings of the IEEE 5th World Forum on Internet of Things (WF-IoT). IEEE, Limerick, Ireland (2019).
Devices on the IoT are usually battery-powered and have limited resources. Hence, energy-efficient and lightweight protocols were designed for IoT devices, such as the popular CoAP. Yet, CoAP itself does not include any defenses against denial-of-sleep attacks, which are attacks that aim at depriving victim devices of entering low-power sleep modes. For example, a denial-of-sleep attack against an IoT device that runs a CoAP server is to send plenty of CoAP messages to it, thereby forcing the IoT device to expend energy for receiving and processing these CoAP messages. All current security solutions for CoAP, namely DTLS, IPsec, and OSCORE, fail to prevent such attacks. To fill this gap, Seitz et al. proposed a method for filtering out inauthentic and replayed CoAP messages "en-route" on 6LoWPAN border routers. In this paper, we expand on Seitz et al.'s proposal in two ways. First, we revise Seitz et al.'s software architecture so that 6LoWPAN border routers can not only check the authenticity and freshness of CoAP messages, but can also perform a wide range of further checks. Second, we propose a couple of such further checks, which, as compared to Seitz et al.'s original checks, more reliably protect IoT devices that run CoAP servers from remote denial-of-sleep attacks, as well as from remote exploits. We prototyped our solution and successfully tested its compatibility with Contiki-NG's CoAP implementation.
35.
Bock, B., Matysik, J.-T., Krentz, K.-F., Meinel, C.: Link Layer Key Revocation and Rekeying for the Adaptive Key Establishment Scheme. Proceedings of the IEEE 5th World Forum on Internet of Things (WF-IoT). IEEE, Limerick, Ireland (2019).
While the IEEE 802.15.4 radio standard has many features that meet the requirements of Internet of things (IoT) applications, IEEE 802.15.4 leaves the whole issue of key management unstandardized. To address this gap, Krentz et al. proposed the Adaptive Key Establishment Scheme (AKES), which establishes session keys for use in IEEE 802.15.4 security. Yet, AKES does not cover all aspects of key management. In particular, AKES comprises no means for key revocation and rekeying. Moreover, existing protocols for key revocation and rekeying seem limited in various ways. In this paper, we hence propose a key revocation and rekeying protocol, which is designed to overcome various limitations of current protocols for key revocation and rekeying. For example, our protocol seems unique in that it routes around IEEE 802.15.4 nodes whose keys are being revoked. We succesfully implemented and evaluated our protocol using the Contiki-NG operating system and aiocoap.
36.
Rohloff, T., Bothe, M., Meinel, C.: Visualizing Content Exploration Traces of MOOC Students. Companion Proceedings of the 9th International Conference on Learning Analytics & Knowledge (LAK 2019). pp. 754–758. SoLAR (2019).
This workshop paper introduces a novel approach to visualize content exploration traces of students who navigate through the learning material of Massive Open Online Courses (MOOCs). This can help teachers to identify trends and anomalies in their provided learning material in order to improve the learning experience. The difficulty lies in the complexity of data: MOOCs are structured into multiple sections consisting of different learning items and students can navigate freely between them. Therefore, it is challenging to find a meaningful and comprehensible visualization that provides a complete overview for teachers. We utilized a Sankey diagram which shows the students' transitions between course sections by grouping them into different buckets, based on the percentage of visited items in the corresponding section. Three preceding data processing steps are explained as well as the data visualization with an example course. This is followed by pedagogical considerations how MOOC teachers can utilize and interpret the visualization, to gain meaningful insights and execute informed actions. At last, an evaluation concept is outlined.
37.
Bethge, J., Yang, H., Bornstein, M., Meinel, C.: BinaryDenseNet: Developing an Architecture for Binary Neural Networks. The IEEE International Conference on Computer Vision (ICCV) Workshops (2019).
38.
Serth, S.: Integrating Professional Tools in Programming Education with MOOCs. Proceedings of the 2019 IEEE Frontiers in Education Conference (FIE). IEEE (2019).
An increasing number of high school teachers use existing Massive Open Online Courses (MOOCs) concerning programming education. Most MOOCs focus on teaching the basics of a programming language and common concepts or patterns. MOOC platforms usually provide their own code execution environments and thus have full control over the features and appearance available to learners. However, only a subset of tools available to professional software engineers is used in introductory programming MOOCs. While the reduction of features is helpful to ease navigation for novices, we assume that learners benefit from more advanced features at a later stage in the learning process. To help students minimize bugs and conceptual mistakes, we intend to evaluate how pair programming could be enabled for remote peers in MOOCs with a synchronized editor and an additional communication channel. Further, we plan to use static program analysis to get more insights about the code written by learners and to provide early feedback about the coding style. One of our contributions will be to identify possibilities to integrate professional tools and methods in MOOCs supported by an evaluation from learners.
39.
Serth, S., Teusner, R., Renz, J., Uflacker, M.: Evaluating Digital Worksheets with Interactive Programming Exercises for K-12 Education. Proceedings of the 2019 IEEE Frontiers in Education Conference (FIE). IEEE (2019).
This Research Full Paper presents insights from digital worksheets with embedded interactive programming exercises tailored for high-school students new to programming. Computer Science teachers often incorporate existing videos, quizzes, and practical programming exercises from Massive Open Online Courses (MOOCs). However, teachers' options to adapt the content to their specific needs are currently limited. Based on a qualitative survey with thirteen teachers, we developed a software prototype which allows teachers to create their own interactive worksheets consisting of texts, videos, quizzes, and practical programming exercises. Additionally, teachers can embed and further customize existing exercises from MOOCs. Further, we enable teachers to gain deeper insights by providing results from automated submission analysis, thus uncovering knowledge gaps and fostering content-driven in-class discussions. Our evaluation shows that the concept was well received by students and teachers alike: Teachers noticed the possibility of a shift in their role from a lecturing instructor to an individual tutor, as students are enabled to learn at their own pace and receive specific, direct feedback based on automated unit tests. Interactive worksheets, as an integrated part of digital education, thus foster informed teacher interventions as part of an individualized student learning process.
40.
Traifeh, H., Staubitz, T., Meinel, C.: Improving learner experience and participation in MOOCs: A design thinking approach. 2019 Learning With MOOCS (LWMOOCS) (2019).
41.
John, C.T., Staubitz, T., Meinel, C.: Performance of Men and Women in Graded Team Assignments in MOOCs. 2019 Learning With MOOCS (LWMOOCS) (2019).
42.
John, C.T., Staubitz, T., Meinel, C.: Took a MOOC. Got a Certificate. What now?. 2019 IEEE Frontiers in Education Conference (FIE) (2019).
43.
Bethge, J., Yang, H., Meinel, C.: Training Accurate Binary Neural Networks from Scratch. 2019 26th IEEE International Conference on Image Processing (ICIP) (2019).
Binary neural networks are a promising approach to execute convolutional neural networks on devices with low computational power. Previous work on this subject often quantizes pretrained full-precision models and uses complex training strategies. In our work, we focus on increasing the performance of binary neural networks by training from scratch with a simple training strategy. In our experiments we show that we are able to achieve state-of-the-art results on standard benchmark datasets. Further, we analyze how full-precision network structures can be adapted for efficient binary networks and adopt a network architecture based on a DenseNet for binary networks, which lets us improve the state-of-the-art even further. Our source code can be found online: https://github.com/hpi-xnor/BMXNet-v2
44.
Serth, S.: Individual Worksheets with Interactive Programming Exercises within the HPI Schul-Cloud. Master’s Thesis. Hasso Plattner Institute, University of Potsdam (2019).
Modern computer science education in high-schools requires students to learn the basics of programming. In terms of content, teachers often incorporate existing videos, quizzes, and practical programming exercises from Massive Open Online Courses (MOOCs). However, teachers' options to adapt the content to their specific needs or to add their own material are currently limited. Based on a qualitative survey with thirteen teachers, we developed tools to extend these options. Our software prototype allows teachers to create their own interactive worksheets consisting of texts, videos, quizzes, and practical programming exercises. Additionally, teachers can embed and further customize existing exercises from MOOCs. Further, we enable teachers to gain deeper insights into the learning progress of their students by providing results from automated submission analysis. These data allow uncovering potential knowledge gaps and foster content-driven in-class discussions. In this thesis, we present findings from an evaluation with different school classes using our software. The concept was well received by students and teachers alike: Teachers noticed the possibility of a shift in their role from a lecturing instructor to an individual tutor, as students are enabled to learn at their own pace and receive specific, direct feedback based on automated unit tests. For the preparation of upcoming lessons, teachers valued the ability to analyze common mistakes of their students to uncover and discuss previously hidden problems. Interactive worksheets, as an integrated part of digital education, thus foster informed teacher interventions as part of an individualized student learning process.
45.
Staubitz, T., Teusner, R., Meinel, C.: MOOCs in Secondary Education - Experiments and Observations from German Classrooms. 2019 IEEE Global Engineering Education Conference (EDUCON). pp. 173–182 (2019).
Computer science education in German schools is often less than optimal. It is only mandatory in a few of the federal states and there is a lack of qualified teachers. As a MOOC (Massive Open Online Course) provider with a German background, we developed the idea to implement a MOOC addressing pupils in secondary schools to fill this gap. The course targeted high school pupils and enabled them to learn the Python programming language. In 2014, we successfully conducted the first iteration of this MOOC with more than 7000 participants. However, the share of pupils in the course was not quite satisfactory. So we conducted several workshops with teachers to find out why they had not used the course to the extent that we had imagined. The paper at hand explores and discusses the steps we have taken in the following years as a result of these workshops.
46.
Bothe, M., Renz, J., Rohloff, T., Meinel, C.: From MOOCs to Micro Learning Activities. 2019 IEEE Global Engineering Education Conference (EDUCON). pp. 280–288 (2019).
Mobile devices are omnipresent in our daily lives. They are utilized for a variety of tasks and used multiple times for short periods throughout the day. MOOC providers optimized their platforms for these devices in order to support ubiquitous learning. While a combination of desktop and mobile learning yields improved course performances, standalone learning on mobile devices does not perform in the same manner. One indicator for this is the mismatch between the average usage pattern of mobile devices and the time to consume one content item in a MOOC. Micro learning builds on bite-sized learning material and focusses on short-term learning sessions. This work examines the potential of micro learning activities in the context of MOOCs. Therefore, a framework for video-based micro learning is presented, which features a personalized curriculum. Videos are suggested to the user in a non-linear order that is determined by content dependencies, users’ preferences and watched videos, as well as explicit and implicit user feedback. A mobile application was implemented to test the approach with restructured MOOC content resulting in 58 connected short videos about engineering education – e.g. web technologies and programming languages. The usage data indicates initial curiosity by the users. To improve retention rates, more user motivation will be required for future studies. A survey gathered additional qualitative feedback. While the content suggestions were seen as a vital feature for such an approach, the results showed good interest and acceptance rates to create a better learning experience for MOOCs on mobile devices.
