Research

Publications

Scheir Peter, Prettenhofer Peter, Lindstaedt Stefanie , Ghidini Chiara
2010
While it is agreed that semantic enrichment of resources would lead to better search results, at present the low coverage of resources on the web with semantic information presents a major hurdle in realizing the vision of search on the Semantic Web. To address this problem, this chapter investigates how to improve retrieval performance in settings where resources are sparsely annotated with semantic information. Techniques from soft computing are employed to find relevant material that was not originally annotated with the concepts used in a query. The authors present an associative retrieval model for the Semantic Web and evaluate if and to which extent the use of associative retrieval techniques increases retrieval performance. In addition, the authors present recent work on adapting the network structure based on relevance feedback by the user to further improve retrieval effectiveness. The evaluation of new retrieval paradigms - such as retrieval based on technology for the Semantic Web - presents an additional challenge since no off-the-shelf test corpora exist. Hence, this chapter gives a detailed description of the approach taken to evaluate the information retrieval service the authors have built.
Balacheff, Nicolas, Bottino, Rosa, Fischer, Frank, Hofmann, Lena, Joubert, Marie, Kieslinger, Barbara, Lindstaedt Stefanie , Manca, Stefanie, Ney, Muriel, Pozzi, Francesca, Sutherland, Rosamund, Verbert, Katrien, Timmis, Sue, Wild, Fridolin, Scott, Peter, Specht, Marcus
2010
This First TEL Grand Challenge Vision and Strategy Report aims to: • provide a unifying framework for members of STELLAR (including doctoral candidates) to develop their own research agenda • engage the STELLAR community in scientific debate and discussion with the long term aim of developing awareness of and respect for different theoretical and methodological perspectives • build knowledge related to the STELLAR grand challenges through the construction of a wiki that is iteratively co‐edited throughout the life of the STELLAR network • develop understandings of the way in which web 2.0 technologies can be used to construct knowledge within a research community (science 2.0) • develop strategies for ways in which the STELLAR instruments can feed into the ongoing development of the wiki and how the they can be used to address the challenges highlighted in this report.
Pozzi, Francesca, Persico, Donatella, Fischer, Frank, Hofmann, Lena, Lindstaedt Stefanie , Cress, Ulrike, Rath Andreas S., Moskaliuk, Johannes, Weber, Nicolas, Kimmerle, Joachim, Devaurs Didier, Ney, Muriel, Gonçalves, Celso, Balacheff, Nicolas, Schwartz, Claudine, Bosson, Jean-Luc, Dillenbourg, Pierre, Jermann, Patrick, Zufferey, Guillaume, Brown, Elisabeth, Sharples, Mike, Windrum, Caroline, Specht, Marcus, Börner, Dirk, Glahn, Christian, Fiedler, Sebastian, Fisichella, Marco, Herder, Eelco, Marenzi, Ivana, Nejdl, Wolfgang, Kawese, Ricardo, Papadakis, George
2010
In this first STELLAR trend report we survey the more distant future of TEL, as reflected in the roadmaps; we compare the visions with trends in TEL research and TEL practice. This generic overview is complemented by a number of small-scale studies, which focus on a specific technology, approach or pedagogical model.
Granitzer Michael, Kienreich Wolfgang, Sabol Vedran, Lex Elisabeth
2010
Technological advances and paradigmatic changes in the utilization of the World Wide Web havetransformed the information seeking strategies of media consumers and invalidated traditionalbusiness models of media providers. We discuss relevant aspects of this development and presenta knowledge relationship discovery pipeline to address the requirements of media providers andmedia consumers. We also propose visually enhanced access methods to bridge the gap betweencomplex media services and the information needs of the general public. We conclude that acombination of advanced processing methods and visualizations will enable media providers totake the step from content-centered to service-centered business models and, at the same time,will help media consumers to better satisfy their personal information needs.
Wolpers Martin, Kirschner Paul A., Scheffel Maren, Lindstaedt Stefanie , Dimitrova Vania
2010
Lindstaedt Stefanie , Duval E., Ullmann T.D., Wild F., Scott P.
2010
Research2.0 is in essence a Web2.0 approach to how we do research. Research2.0 creates conversations between researchers, enables them to discuss their findings and connects them with others. Thus, Research2.0 can accelerate the diffusion of knowledge.ChallengesAs concluded during the workshop, at least four challenges are vital for future research.The first area is concerned with availability of data. Access to sanitized data and conventions on how to describe publication-related metadata provided from divergent sources are enablers for researchers to develop new views on their publications and their research area. Additional, social media data gain more and more attention. Reaching a widespread agreement about this for the field of technology-enhanced learning would be already a major step, but it is also important to focus on the next steps: what are success-critical added values driving uptake in the research community as a whole?The second area of challenges is seen in Research 2.0 practices. As technology-enhanced learning is a multidisciplinary field, practices developed in one area could be valuable for others. To extract the essence of successful multidisciplinary Research 2.0 practice though, multidimensional and longitudinal empirical work is needed. It is also an open question, if we should support practice by fostering the usage of existing tools or the development of new tools, which follow Research 2.0 principles. What makes a practice sustainable? What are the driving factors?The third challenge deals with impact. What are criteria of impact for research results (and other research artefacts) published on the Web? How can this be related to the publishing world appearing in print? Is a link equal to a citation or a download equal to a subscription? Can we develop a Research 2.0 specific position on impact measurement? This includes questions of authority, quality and re-evaluation of quality, and trust.The tension between openness and privacy spans the fourth challenge. The functionality of mash-ups often relies on the use of third-party services. What happens with the data, if this source is no longer available? What about hidden exchange of data among backend services?
Lindstaedt Stefanie , Rath Andreas S., Devaurs Didier
2010
. Supporting learning activities during work has gained momentum fororganizations since work-integrated learning (WIL) has been shown to increaseproductivity of knowledge workers. WIL aims at fostering learning at the workplace,during work, for enhancing task performance. A key challenge for enablingtask-specific, contextualized, personalized learning and work support is to automaticallydetect the user’s task. In this paper we utilize our ontology-based usertask detection approach for studying the factors influencing task detection performance.We describe three laboratory experiments we have performed in twodomains including over 40 users and more than 500 recorded task executions.The insights gained from our evaluation are: (i) the J48 decision tree and Na¨ıveBayes classifiers perform best, (ii) six features can be isolated, which providegood classification accuracy, (iii) knowledge-intensive tasks can be classified aswell as routine tasks and (iv) a classifier trained by experts on standardized taskscan be used to classify users’ personal tasks.
Kern Roman, Granitzer Michael, Muhr M.
2010
Word sense induction and discrimination(WSID) identifies the senses of an ambiguousword and assigns instances of thisword to one of these senses. We have builda WSID system that exploits syntactic andsemantic features based on the results ofa natural language parser component. Toachieve high robustness and good generalizationcapabilities, we designed our systemto work on a restricted, but grammaticallyrich set of features. Based on theresults of the evaluations our system providesa promising performance and robustness.
Granitzer Michael, Sabol Vedran, Onn K., Lukose D.
2010
Schachner W.
2010
Schachner W.
2010
Schachner W.
2010
Stern Hermann, Kaiser Rene_DB, Hofmair P., Lindstaedt Stefanie , Scheir Peter, Kraker Peter
2010
One of the success factors of Work Integrated Learning (WIL) is to provide theappropriate content to the users, both suitable for the topics they are currently working on, andtheir experience level in these topics. Our main contributions in this paper are (i) overcomingthe problem of sparse content annotation by using a network based recommendation approachcalled Associative Network, which exploits the user context as input; (ii) using snippets for notonly highlighting relevant parts of documents, but also serving as a basic concept enabling theWIL system to handle text-based and audiovisual content the same way; and (iii) using the WebTool for Ontology Evaluation (WTE) toolkit for finding the best default semantic similaritymeasure of the Associative Network for new domains. The approach presented is employed inthe software platform APOSDLE, which is designed to enable knowledge workers to learn atwork.
Lindstaedt Stefanie , Kump Barbara, Beham Günter, Pammer-Schindler Viktoria, Ley Tobias, de Hoog R., Dotan A.
2010
We present a work-integrated learning (WIL) concept which aims atempowering employees to learn while performing their work tasks. Withinthree usage scenarios we introduce the APOSDLE environment whichembodies the WIL concept and helps knowledge workers move fluidly alongthe whole spectrum of WIL activities. By doing so, they are experiencingvarying degrees of learning guidance: from building awareness, over exposingknowledge structures and contextualizing cooperation, to triggering reflectionand systematic competence development. Four key APOSDLE components areresponsible for providing this variety of learning guidance. The challenge intheir design lies in offering learning guidance without being domain-specificand without relying on manually created learning content. Our three monthsummative workplace evaluation within three application organizationssuggests that learners prefer awarenss building functionalities and descriptivelearning guidance and reveals that they benefited from it.
Lindstaedt Stefanie , Beham Günter, Stern Hermann, Drachsler H., Bogers T., Vuorikari R., Verbert K., Duval E., Manouselis N., Friedrich M., Wolpers M.
2010
This paper raises the issue of missing data sets for recommender systems in Technology Enhanced Learning that can be used asbenchmarks to compare different recommendation approaches. It discusses how suitable data sets could be created according tosome initial suggestions, and investigates a number of steps that may be followed in order to develop reference data sets that willbe adopted and reused within a scientific community. In addition, policies are discussed that are needed to enhance sharing ofdata sets by taking into account legal protection rights. Finally, an initial elaboration of a representation and exchange format forsharable TEL data sets is carried out. The paper concludes with future research needs.
Beham Günter, Kump Barbara, Lindstaedt Stefanie , Ley Tobias
2010
According to studies into learning at work, interpersonal help seeking is the most important strategy of how people acquireknowledge at their workplaces. Finding knowledgeable persons, however, can often be difficult for several reasons. Expertfinding systems can support the process of identifying knowledgeable colleagues thus facilitating communication andcollaboration within an organization. In order to provide the expert finding functionality, an underlying user model is needed thatrepresents the characteristics of each individual user. In our article we discuss requirements for user models for the workintegratedlearning (WIL) situation. Then, we present the APOSDLE People Recommender Service which is based on anunderlying domain model, and on the APOSDLE User Model. We describe the APOSDLE People Recommender Service on thebasis of the Intuitive Domain Model of expert finding systems, and explain how this service can support interpersonal helpseeking at workplaces.
Lindstaedt Stefanie , Kraker Peter, Höfler Patrick, Fessl Angela
2010
In this paper we present an ecosystem for the lightweight exchangeof publication metadata based on the principles of Web 2.0. At the heart of thisecosystem, semantically enriched RSS feeds are used for dissemination. Thesefeeds are complemented by services for creation and aggregation, as well aswidgets for retrieval and visualization of publication metadata. In twoscenarios, we show how these publication feeds can benefit institutions,researchers, and the TEL community. We then present the formats, services,and widgets developed for the bootstrapping of the ecosystem. We concludewith an outline of the integration of publication feeds with the STELLARNetwork of Excellence1 and an outlook on future developments.
Beham Günter, Lindstaedt Stefanie , Ley Tobias, Kump Barbara, Seifert C.
2010
When inferring a user’s knowledge state from naturally occurringinteractions in adaptive learning systems, one has to makes complexassumptions that may be hard to understand for users. We suggestMyExperiences, an open learner model designed for these specificrequirements. MyExperiences is based on some of the key design principles ofinformation visualization to help users understand the complex information inthe learner model. It further allows users to edit their learner models in order toimprove the accuracy of the information represented there.
Ley Tobias, Seitlinger Paul
2010
Researching the emergence of semantics in social systems needs totake into account how users process information in their cognitive system. Wereport results of an experimental study in which we examined the interactionbetween individual expertise and the basic level advantage in collaborative tagging.The basic level advantage describes availability in memory of certain preferredlevels of taxonomic abstraction when categorizing objects and has beenshown to vary with level of expertise. In the study, groups of students taggedinternet resources for a 10-week period. We measured the availability of tags inmemory with an association test and a relevance rating and found a basic leveladvantage for tags from more general as opposed to specific levels of the taxonomy.An interaction with expertise also emerged. Contrary to our expectations,groups that spent less time to develop a shared understanding shifted tomore specific levels as compared to groups that spent more time on a topic. Weattribute this to impaired collaboration in the groups. We discuss implicationsfor personalized tag and resource recommendations.
Beham Günter, Jeanquartier Fleur, Lindstaedt Stefanie
2010
This paper introduces iAPOSDLE, a mobile application enabling the use of work-integrated learning services without being limited by location. iAPOSDLE makes use of the APOSDLE WIL system for self-directed work-integrated learning support, and extends its range of application to mobile learning. Core features of iAPOSDLE are described and possible extensions are discussed.
Kern Roman, Granitzer Michael, Muhr M.
2010
Cluster label quality is crucial for browsing topic hierarchiesobtained via document clustering. Intuitively, the hierarchicalstructure should influence the labeling accuracy. However,most labeling algorithms ignore such structural propertiesand therefore, the impact of hierarchical structureson the labeling accuracy is yet unclear. In our work weintegrate hierarchical information, i.e. sibling and parentchildrelations, in the cluster labeling process. We adaptstandard labeling approaches, namely Maximum Term Frequency,Jensen-Shannon Divergence, χ2 Test, and InformationGain, to take use of those relationships and evaluatetheir impact on 4 different datasets, namely the Open DirectoryProject, Wikipedia, TREC Ohsumed and the CLEFIP European Patent dataset. We show, that hierarchicalrelationships can be exploited to increase labeling accuracyespecially on high-level nodes.
Lex Elisabeth, Granitzer Michael, Juffinger A.
2010
In the blogosphere, the amount of digital content is expanding and for search engines, new challenges have been imposed. Due to the changing information need, automatic methods are needed to support blog search users to filter information by different facets. In our work, we aim to support blog search with genre and facet information. Since we focus on the news genre, our approach is to classify blogs into news versus rest. Also, we assess the emotionality facet in news related blogs to enable users to identify people’s feelings towards specific events. Our approach is to evaluate the performance of text classifiers with lexical and stylometric features to determine the best performing combination for our tasks. Our experiments on a subset of the TREC Blogs08 dataset reveal that classifiers trained on lexical features perform consistently better than classifiers trained on the best stylometric features.
Kröll Mark, Strohmaier M.
2010
In this paper, we introduce the idea of Intent Analysis, which is to create a profile of the goals and intentions present in textual content. Intent Analysis, similar to Sentiment Analysis, represents a type of document classification that differs from traditional topic categorization by focusing on classification by intent. We investigate the extent to which the automatic analysis of human intentions in text is feasible and report our preliminary results, and discuss potential applications. Inaddition, we present results from a study that focused on evaluating intent profiles generated from transcripts of American presidential candidate speeches in 2008.
Stocker A., Mueller J.
2010
Ley Tobias, Kump Barbara, Gerdenitsch C.
2010
Adaptive scaffolding has been proposed as an efficient means for supporting self-directed learning both in educational as well as in adaptive learning systems research. However, the effects of adaptation on self-directed learning and the differential contributions of different adaptation models have not been systematically examined. In this paper, we examine whether personalized scaffolding in the learning process improves learning. We conducted a controlled lab study in which 29 students had to solve several tasks and learn with the help of an adaptive learning system in a within-subjects control condition design. In the learning process, participants obtained recommendations for learning goals from the system in three conditions: fixed scaffolding where learning goals were generated from the domain model, personalized scaffolding where these recommendations were ranked according to the user model, and random suggestions of learning goals (control condition). Students in the two experimental conditions clearly outperformed students in the control condition and felt better supported by the system. Additionally, students who received personalized scaffolding selected fewer learning goals than participants from the other groups.
Lex Elisabeth, Granitzer Michael, Juffinger A.
2010
In this paper, we outline our experiments carried out at the TREC 2009 Blog Distillation Task. Our system is based on a plain text index extracted from the XML feeds of the TREC Blogs08 dataset. This index was used to retrieve candidate blogs for the given topics. The resulting blogs were classified using a Support Vector Machine that was trained on a manually labelled subset of the TREC Blogs08 dataset. Our experiments included three runs on different features: firstly on nouns, secondly on stylometric properties, and thirdly on punctuation statistics. The facet identification based on our approach was successful, although a significant number of candidate blogs were not retrieved at all.
Granitzer Michael, Kienreich Wolfgang
2010
Granitzer Michael
2010
Term weighting strongly influences the performance of text miningand information retrieval approaches. Usually term weights are determined throughstatistical estimates based on static weighting schemes. Such static approacheslack the capability to generalize to different domains and different data sets. Inthis paper, we introduce an on-line learning method for adapting term weightsin a supervised manner. Via stochastic optimization we determine a linear transformationof the term space to approximate expected similarity values amongdocuments. We evaluate our approach on 18 standard text data sets and showthat the performance improvement of a k-NN classifier ranges between 1% and12% by using adaptive term weighting as preprocessing step. Further, we provideempirical evidence that our approach is efficient to cope with larger problems
Ley Tobias, Kump Barbara, Albert D.
2010