Know Center Research

Publikationen

Gursch Heimo, Schlager Elke, Feichtinger Gerald, Brandl Daniel
2020
The comfort humans perceive in rooms depends on many influencing factors and is currently only poorly recorded and maintained. This is due to circumstances like the subjective nature of perceived comfort, lack of sensors or data processing infrastructure. Project COMFORT (Comfort Orientated and Management Focused Operation of Room condiTions) researches the modelling of perceived thermal comfort of humans in office rooms. This begins at extensive and long-term measurements taking in a laboratory test chamber and in real-world office rooms. Data is collected from the installed building services engineering systems, from high-accurate reference measurement equipment and from weather services describing the outside conditions. All data is stored in a specially developed central Data Management System (DMS) creating the basis for all research and studies in project COMFORT.The collected data is the key enabler for the creation of soft sensors describing comfort relevant indices like predicted mean vote (PMV), predicted percentage of dissatisfied (PPD) and operative temperature (OT). Two different approaches are conducted complementing and extending each other in the realisation of soft sensors. Firstly, a purely data-driven modelling approach generates models for soft sensors by learning the relations between explanatory and target variables in the collected data. Secondly, simulation-based soft sensors are derived from Building Energy Simulation (BES) and Computational Fluid Dynamic (CFD) simulations.The first result of the data-driven analysis is a solar Radiation Modelling (RM) component, capable of splitting global radiation into its direct horizontal and diffuse components. This is needed, since only global radiation data is available for the investigated locations, but the global radiation needs to be divided into direct and diffuse radiation due to their hugely differences in their thermal impact on buildings. The current BES and CFD simulation provide as their results soft sensors for comfort relevant indices, which will be complemented by data-driven soft sensors in the remainder of the project.
Barreiros Carla, Silva Nelson, Veas Eduardo Enrique, Pammer-Schindler Viktoria
2020
Duricic Tomislav, Hussain Hussain, Lacic Emanuel, Kowald Dominik, Lex Elisabeth, Helic Denis
2020
In this work, we study the utility of graph embeddings to generate latent user representations for trust-based collaborative filtering. In a cold-start setting, on three publicly available datasets, we evaluate approaches from four method families:(i) factorization-based,(ii) random walk-based,(iii) deep learning-based, and (iv) the Large-scale Information Network Embedding (LINE) approach. We find that across the four families, random-walk-based approaches consistently achieve the best accuracy. Besides, they result in highly novel and diverse recommendations. Furthermore, our results show that the use of graph embeddings in trust-based collaborative filtering significantly improves user coverage
Reiter-Haas Markus, Wittenbrink Davi, Lacic Emanuel
2020
Finding the right job is a difficult task for anyone as it usually depends on many factors like salary, job description, or geographical location. Students with almost no prior experience, especially, have a hard time on the job market, which is very competitive in nature. Additionally, students often suffer a lack of orientation, as they do not know what kind of job is suitable for their education. At Talto1, we realized this and have built a platform to help Austrian university students with finding their career paths as well as providing them with content that is relevant to their career possibilities. This is mainly achieved by guiding the students toward different types of entities that are related to their career, i.e., job postings, company profiles, and career-related articles.In this talk, we share our experiences with solving the recommendation problem for university students. One trait of the student-focused job domain is that behaviour of the students differs depending on their study progression. At the beginning of their studies, they need study-specific career information and part-time jobs to earn additional money. Whereas, when they are nearing graduation, they require information about their potential future employers and entry-level full-time jobs. Moreover, we can observe seasonal patterns in user activity in addition to the need of handling both logged-in and anonymous session users at the same time.To cope with the requirements of the job domain, we built hybrid models based on a microservice architecture that utilizes popular algorithms from the literature such as Collaborative Filtering, Content-based Filtering as well as various neural embedding approaches (e.g., Doc2Vec, Autoencoders, etc.). We further adapted our architecture to calculate relevant recommendations in real-time (i.e., after a recommendation is requested) as individual user sessions in Talto are usually short-lived and context-dependent. Here we found that the online performance of the utilized approach also depends on the location context [1]. Hence, the current location of a user on the mobile or web application impacts the expected recommendations.One optimization criterion on the Talto career platform is to provide relevant cross-entity recommendations as well as explain why those were shown. Recently, we started to tackle this by learning embeddings of entities that lie in the same embedding space [2]. Specifically, we pre-train word embeddings and link different entities by shared concepts, which we use for training the network embeddings. This embeds both the concepts and the entities into a common vector space, where the common vector space is a result of considering the textual content, as well as the network information (i.e., links to concepts). This way, different entity types (e.g., job postings, company profiles, and articles) are directly comparable and are suited for a real-time recommendation setting. Interestingly enough, with such an approach we also end up with individual words sharing the same embedding space. This, in turn, can be leveraged to enhance the textual search functionality of a platform, which is most commonly based just on a TF-IDF model.Furthermore, we found that such embeddings allow us to tackle the problem of explainability in an algorithm-agnostic way. Since the Talto platform utilizes various recommendation algorithms as well as continuously conducts AB tests, an algorithm-agnostic explainability model would be best suited to provide the students with meaningful explanations. As such, we will also go into the details on how we can adapt our explanation model to not rely on the utilized recommendation algorithm.
Dennerlein Sebastian, Wolf-Brenner Christof, Gutounig Robert, Schweiger Stefan, Pammer-Schindler Viktoria
2020
In society and politics, there is a rising interest in considering ethical principles in technological innovation, especially in the intersection of education and technology. We propose a first iteration of a theory-derived framework to analyze ethical issues in technology-enhanced learning (TEL) software development. The framework understands ethical issues as an expression of the overall socio-technical system that are rooted in the interactions of human actors with technology, so-called socio-technical interactions (STIs). For guiding ethical reflection, the framework helps to explicate this human involvement, and to elicit discussions of ethical principles on these STIs. Prompts in the form of reflection questions can be inferred to reflect on the technology functionality from relevant human perspectives, and in relation to a list of fundamental ethical principles. We illustrate the framework and discuss its implications for TEL
Gayane Sedrakya, Dennerlein Sebastian, Pammer-Schindler Viktoria, Lindstaedt Stefanie
2020
Our earlier research attempts to close the gap between learning behavior analytics based dashboard feedback and learning theories by grounding the idea of dashboard feedback onto learning science concepts such as feedback, learning goals, (socio-/meta-) cognitive mechanisms underlying learning processes. This work extends the earlier research by proposing mechanisms for making those concepts and relationships measurable. The outcome is a complementary framework that allows identifying feedback needs and timing for their provision in a generic context that can be applied to a certain subject in a given LMS. The research serves as general guidelines for educators in designing educational dashboards, as well as a starting research platform in the direction of systematically matching learning sciences concepts with data and analytics concepts
Klimashevskaia Anastasia, Geiger Bernhard, Hagmüller Martin, Helic Denis, Fischer Frank
2020
(extended abstract)
Hobisch Elisbeth, Scholger Martina, Fuchs Alexandra, Geiger Bernhard, Koncar Philipp, Saric Sanja
2020
(extended abstract)
Schrunner Stefan, Geiger Bernhard, Zernig Anja, Kern Roman
2020
Classification has been tackled by a large number of algorithms, predominantly following a supervised learning setting. Surprisingly little research has been devoted to the problem setting where a dataset is only partially labeled, including even instances of entirely unlabeled classes. Algorithmic solutions that are suited for such problems are especially important in practical scenarios, where the labelling of data is prohibitively expensive, or the understanding of the data is lacking, including cases, where only a subset of the classes is known. We present a generative method to address the problem of semi-supervised classification with unknown classes, whereby we follow a Bayesian perspective. In detail, we apply a two-step procedure based on Bayesian classifiers and exploit information from both a small set of labeled data in combination with a larger set of unlabeled training data, allowing that the labeled dataset does not contain samples from all present classes. This represents a common practical application setup, where the labeled training set is not exhaustive. We show in a series of experiments that our approach outperforms state-of-the-art methods tackling similar semi-supervised learning problems. Since our approach yields a generative model, which aids the understanding of the data, it is particularly suited for practical applications.
Gogolenko Sergiy, Groen Derek, Suleimenova Dian, Mahmood Imra, Lawenda Marcin, Nieto De Santos Javie, Hanley Joh, Vukovic Milana, Kröll Mark, Geiger Bernhard, Elsaesser Rober, Hoppe Dennis
2020
Accurate digital twinning of the global challenges (GC) leadsto computationally expensive coupled simulations. These simulationsbring together not only different models, but also various sources of mas-sive static and streaming data sets. In this paper, we explore ways tobridge the gap between traditional high performance computing (HPC)and data-centric computation in order to provide efficient technologicalsolutions for accurate policy-making in the domain of GC. GC simula-tions in HPC environments give rise to a number of technical challengesrelated to coupling. Being intended to reflect current and upcoming situ-ation for policy-making, GC simulations extensively use recent streamingdata coming from external data sources, which requires changing tradi-tional HPC systems operation. Another common challenge stems fromthe necessity to couple simulations and exchange data across data centersin GC scenarios. By introducing a generalized GC simulation workflow,this paper shows commonality of the technical challenges for various GCand reflects on the approaches to tackle these technical challenges in theHiDALGO project
Breitfuß Gert, Fruhwirth Michael, Wolf-Brenner Christof, Riedl Angelika, Ginthör Robert, Pimas Oliver
2020
In the future, every successful company must have a clear idea of what data means to it. The necessary transformation to a data-driven company places high demands on companies and challenges management, organization and individual employees. In order to generate concrete added value from data, the collaboration of different disciplines e.g. data scientists, domain experts and business people is necessary. So far few tools are available which facilitate the creativity and co-creation process amongst teams with different backgrounds. The goal of this paper is to design and develop a hands-on and easy to use card-based tool for the generation of data service ideas that supports the required interdisciplinary cooperation. By using a Design Science Research approach we analysed 122 data service ideas and developed an innovation tool consisting of 38 cards. The first evaluation results show that the developed Data Service Cards are both perceived as helpful and easy to use.
Fruhwirth Michael, Breitfuß Gert, Pammer-Schindler Viktoria
2020
The availability of data sources and advances in analytics and artificial intelligence offers the opportunity for organizationsto develop new data-driven products, services and business models. Though, this process is challenging for traditionalorganizations, as it requires knowledge and collaboration from several disciplines such as data science, domain experts, orbusiness perspective. Furthermore, it is challenging to craft a meaningful value proposition based on data; whereas existingresearch can provide little guidance. To overcome those challenges, we conducted a Design Science Research project toderive requirements from literature and a case study, develop a collaborative visual tool and evaluate it through severalworkshops with traditional organizations. This paper presents the Data Product Canvas, a tool connecting data sources withthe user challenges and wishes through several intermediate steps. Thus, this paper contributes to the scientific body ofknowledge on developing data-driven business models, products and services.
Dumouchel Suzanne, Blotiere Emilie, Barbot Laure, Breitfuß Gert, Chen Yin, Di Donato Francesca, Forbes Paula, Petifils Clara, Pohle Stefanie
2020
SSH research is divided across a wide array of disciplines, sub-disciplines, and languages. While this specialisation makes it possible to investigate the extensive variety of SSH topics, it also leads to a fragmentation that prevents SSH research from reaching its full potential. Use and reuse of SSH research is suboptimal, interdisciplinary collaboration possibilities are often missed partially because of missing standards and referential keys between disciplines. By the way the reuse of data may paradoxically complicate a relevant sorting and a trust relationship. As a result, societal, economic and academic impacts are limited. Conceptually, there is a wealth of transdisciplinary collaborations, but in practice there is a need to help SSH researchers and research institutions to connect them and support them, to prepare the research data for these overarching approaches and to make them findable and usable. The TRIPLE (Targeting Researchers through Innovative Practices and Linked Exploration) project is a practical answer to the above issues, as it aims at designing and developing the European discovery platform dedicated to SSH resources. Funded under the European Commission program INFRAEOSC-02-2019 “Prototyping new innovative services”, thanks to a consortium of 18 partners, TRIPLE will develop a full multilingual and multicultural solution for the discovery and the reuse of SSH resources. The project started in October 2019 for a duration of 42 months thanks to European funding of 5.6 million €.
Kowald Dominik, Markus Schedl, Lex Elisabeth
2020
Research has shown that recommender systems are typicallybiased towards popular items, which leads to less popular items beingunderrepresented in recommendations. The recent work of Abdollahpouriet al. in the context of movie recommendations has shown that this pop-ularity bias leads to unfair treatment of both long-tail items as well asusers with little interest in popular items. In this paper, we reproducethe analyses of Abdollahpouri et al. in the context of music recommen-dation. Specifically, we investigate three user groups from the Last.fmmusic platform that are categorized based on how much their listen-ing preferences deviate from the most popular music among all Last.fmusers in the dataset: (i) low-mainstream users, (ii) medium-mainstreamusers, and (iii) high-mainstream users. In line with Abdollahpouri et al.,we find that state-of-the-art recommendation algorithms favor popularitems also in the music domain. However, their proposed Group Aver-age Popularity metric yields different results for Last.fm than for themovie domain, presumably due to the larger number of available items(i.e., music artists) in the Last.fm dataset we use. Finally, we comparethe accuracy results of the recommendation algorithms for the three usergroups and find that the low-mainstreaminess group significantly receivesthe worst recommendations.
Dennerlein Sebastian, Pammer-Schindler Viktoria, Ebner Markus, Getzinger Günter, Ebner Martin
2020
Sustainably digitalizing higher education requires a human-centred approach. To address actual problems in teaching as well as learning and increase acceptance, the Technology Enhanced Learning (TEL) solution(s) must be co-designed with affected researchers, teachers, students and administrative staff. We present research-in-progress about a sandpit-informed innovation process with a f2f-marketplace of TEL research and problemmapping as well team formation alongside a competitive call phase, which is followed by a cooperative phase of funded interdisciplinary pilot teams codesigning and implementing TEL innovations. Pilot teams are supported by a University Innovation Canvas to document and reflect on their TEL innovation from multiple viewpoints.
Fuchs Alexandra, Geiger Bernhard, Hobisch Elisabeth, Koncar Philipp, More Jacqueline, Saric Sanja, Scholger Martina
2020
Feichtinger Gerald, Gursch Heimo, Schlager Elke, Brandl Daniel, Gratzl Markus
2020
Fadljevic Leon, Maitz Katharina, Kowald Dominik, Pammer-Schindler Viktoria, Gasteiger-Klicpera Barbara
2020
This paper describes the analysis of temporal behavior of 11--15 year old students in a heavily instructionally designed adaptive e-learning environment. The e-learning system is designed to support student's acquisition of health literacy. The system adapts text difficulty depending on students' reading competence, grouping students into four competence levels. Content for the four levels of reading competence was created by clinical psychologists, pedagogues and medicine students. The e-learning system consists of an initial reading competence assessment, texts about health issues, and learning tasks related to these texts. The research question we investigate in this work is whether temporal behavior is a differentiator between students despite the system's adaptation to students' reading competence, and despite students having comparatively little freedom of action within the system. Further, we also investigated the correlation of temporal behaviour with performance. Unsupervised clustering clearly separates students into slow and fast students with respect to the time they take to complete tasks. Furthermore, topic completion time is linearly correlated with performance in the tasks. This means that we interpret working slowly in this case as diligence, which leads to more correct answers, even though the level of text difficulty matches student's reading competence. This result also points to the design opportunity to integrate advice on overarching learning strategies, such as working diligently instead of rushing through, into the student's overall learning activity. This can be done either by teachers, or via additional adaptive learning guidance within the system.
Thalmann Stefan, Fessl Angela, Pammer-Schindler Viktoria
2020
Digitization is currently one of the major factors changing society and the business world. Most research focused on the technical issues of this change, but also employees and especially the way how they learn changes dramatically. In this paper, we are interested in exploring the perspectives of decision makers in huge manufacturing companies on current challenges in organizing learning and knowledge distribution in digitized manufacturing environments. Moreover, weinvestigated the change process and challenges of implementing new knowledge and learning processes.To this purpose, we have conducted 24 interviews with senior representatives of large manufacturing companies from Austria, Germany, Italy, Liechtenstein and Switzerland. Our exploratory study shows that decision makers perceive significant changes in work practice of manufacturing due to digitization and they currently plan changes in organizational training and knowledge distribution processes in response. Due to the lack of best practices, companies focus verymuch on technological advancements. The delivery of knowledge just-in-time directly into work practice is afavorite approach. Overall, digital learning services are growing and new requirements regarding compliance,quality management and organisational culture arise.
Fruhwirth Michael, Rachinger Michael, Prlja Emina
2020
The modern economy relies heavily on data as a resource for advancement and growth. Data marketplaces have gained an increasing amount of attention, since they provide possibilities to exchange, trade and access data across organizations. Due to the rapid development of the field, the research on business models of data marketplaces is fragmented. We aimed to address this issue in this article by identifying the dimensions and characteristics of data marketplaces from a business model perspective. Following a rigorous process for taxonomy building, we propose a business model taxonomy for data marketplaces. Using evidence collected from a final sample of twenty data marketplaces, we analyze the frequency of specific characteristics of data marketplaces. In addition, we identify four data marketplace business model archetypes. The findings reveal the impact of the structure of data marketplaces as well as the relevance of anonymity and encryption for identified data marketplace archetypes.
Kaiser Rene_DB, Thalmann Stefan, Pammer-Schindler Viktoria, Fessl Angela
2020
Organisations participate in collaborative projects that include competitors for a number of strategic reasons, even whilst knowing that this requires them to consider both knowledge sharing and knowledge protection throughout collaboration. In this paper, we investigated which knowledge protection practices representatives of organizations employ in a collaborative research and innovation project that can be characterized as a co-opetitive setting. We conducted a series of 30 interviews and report the following seven practices in structured form: restrictive partner selection in operative project tasks, communication through a gatekeeper, to limit access to a central platform, to hide details of machine data dumps, to have data not leave a factory for analysis, a generic model enabling to hide usage parameters, and to apply legal measures. When connecting each practice to a priori literature, we find three practices focussing on collaborative data analytics tasks had not yet been covered so far.