Autor*innen mit den meisten Dokumenten
Auflistung nach:
Neueste Veröffentlichungen
- ZeitschriftenartikelThe German EU Council Presidency Translator(KI - Künstliche Intelligenz: Vol. 36, No. 1, 2022) Pinnis, Mārcis; Busemann, Stephan; Vasiļevskis, Artūrs; Genabith, JosefThis contribution describes the German EU Council Presidency Translator (EUC PT), a machine translation service created for the German EU Council Presidency in the second half of 2020, which is open to the general public. Following a series of earlier presidency translators, the German version exhibits important extensions and improvements. The German EUC PT is the first to integrate systems from commercial vendors, public services, and a research center, using a mix of custom and generic translation engines, and to introduce a new webpage translation widget. A further important feature is the close collaboration with human translators from the German ministries, who were provided with computer-assisted translation tool plugins integrating machine translation services into their daily work environments. Uptake by the public reflects a huge interest in the service, showing the need for breaking language barriers.
- ZeitschriftenartikelExpertise depends on reasoning through alternative scenarios(KI - Künstliche Intelligenz: Vol. 36, No. 1, 2022) Dohn, Nina Bonderup; Ragni, Marco
- ZeitschriftenartikelMulti-phase Fine-Tuning: A New Fine-Tuning Approach for Sign Language Recognition(KI - Künstliche Intelligenz: Vol. 36, No. 1, 2022) Sarhan, Noha; Lauri, Mikko; Frintrop, SimoneIn this paper, we propose multi-phase fine-tuning for tuning deep networks from typical object recognition to sign language recognition (SLR). It extends the successful idea of transfer learning by fine-tuning the network’s weights over several phases. Starting from the top of the network, layers are trained in phases by successively unfreezing layers for training. We apply this novel training approach to SLR, since in this application, training data is scarce and differs considerably from the datasets which are usually used for pre-training. Our experiments show that multi-phase fine-tuning can reach significantly better accuracy in fewer training epochs compared to previous fine-tuning techniques
- ZeitschriftenartikelProgramming and Computational Thinking in Mathematics Education(KI - Künstliche Intelligenz: Vol. 36, No. 1, 2022) Tamborg, Andreas Lindenskov; Elicer, Raimundo; Spikol, DanielArtificial intelligence (AI) has become a part of everyday interactions with pervasive digital systems. This development increasingly calls for citizens to have a basic understanding of programming and computational thinking (PCT). Accordingly, countries worldwide are implementing several approaches to integrate critical elements of PCT into K-9 education. However, these efforts are confronted by difficulties that the PCT concepts are for students to grasp from purely theoretical perspectives. Recent literature indicates that the playful nature is particularly important when novices from both both early and higher education are to learn AI. These playful activities are characterised by setting a scene where PCT concepts such as algorithms, data processing, and simulations are meant to draw on to understand better how AI is integrated into our everyday digital life. This discussion paper analyses playful PCT resources developed around the game rock-paper-scissors developed in the UK and Denmark. Resources from these countries are interesting starting points since both have been or are in the process of integrating PCT as part of the K-9 curriculum. The central discussion raised by the paper, is the nature of the integration between mathematics and PCT in these tasks. These resources provide opportunities for discussion of how we may better integrate PCT and mathematics from the perspective of both subjects to build a solid foundation for a critical understanding of AI interactions in future generations.
- ZeitschriftenartikelNews(KI - Künstliche Intelligenz: Vol. 36, No. 1, 2022)
- ZeitschriftenartikelNews(KI - Künstliche Intelligenz: Vol. 36, No. 1, 2022)
- ZeitschriftenartikelFabric-Based Computing: (Re)examining the Materiality of Computer Science Learning Through Fiber Crafts(KI - Künstliche Intelligenz: Vol. 36, No. 1, 2022) Keune, AnnaFiber crafts, such as weaving and sewing, occupy a tension-filled space within computing. While associated with domestic practices, fiber crafts have been recognized as a precursor of the earliest computers and continue to present sources of computational inspiration. The connections between fiber crafts and computing have the potential to uncover possibilities for computing to become more diversified in terms of materials, cultural practices, and ultimately people. To explore the promises of fiber crafts for STEM education, this qualitative dissertation built on constructionist and posthumanist perspectives to examine two fiber crafts (i.e., weaving and fabric manipulation) as contexts for computer science learning. Collectively, the dissertation effectively aligned fiber crafts with computational concepts and showed their potential as a promising context for computer science learning. The work further showed that materials used for STEM learning are non-neutral. Materials matter in what can be learned computationally. Lastly, guided by posthumanist perspectives, the dissertation uncovered computational learning as the process of producing physical expansions and highlighted learning as the process of how computational concepts physically change. The work has implications for theorizing learning, designing for learning, and educational practice. For example, the dissertation presents the utility of posthumanist perspectives as an additional theoretical approach to the study of learning that can surface and help address ongoing relational deficit orientations.
- ZeitschriftenartikelSimplifying Programming for Non-technical Students: A Hermeneutic Approach(KI - Künstliche Intelligenz: Vol. 36, No. 1, 2022) Valente, Andrea; Marchetti, EmanuelaThis paper investigates the simplification of programming for non-technical university students. Typical simplification strategies are outlined, and according to our findings CT courses for non-technical students typically address learners from different faculties, providing generic and basic knowledge, not specifically related to their major. In this study, we propose instead a hermeneutic approach to simplify programming, in which we aim at clarifying the problem-solving aspect of programming, addressing computational problems that are specific to their studies and leveraging on learners’ preunderstanding of the digital media they have experienced as users. The practical counterpart of our theoretical approach is a minimalistic Python multimedia library, called Medialib, that we designed to enable university students with a non-technical profile to create visual media and games with short and readable code. We discuss the use of Medialib in two empirical case studies: a collaboration with the university of Kyushu in Fukuoka, Japan, and a coding module for Media Studies students at the University of Southern Denmark. Furthermore, we use Notional Machines to attempt a comparison of the simplicity of learning tools for programming, and to ground our claim that Medialib is “simpler” for learners than other popular approaches. The main contribution is a hermeneutic approach to the simplification of programming for specific contexts that combines the hermeneutic spiral and notional machines. The approach is supported by a tool, the Medialib library; the two case studies provide examples of how the approach and tool can be deployed in beginners in CT courses.
- ZeitschriftenartikelPrimary Mathematics Teachers’ Understanding of Computational Thinking(KI - Künstliche Intelligenz: Vol. 36, No. 1, 2022) Nordby, Siri Krogh; Bjerke, Annette Hessen; Mifsud, LouiseComputational thinking (CT) is often regarded as providing a ‘soft start’ for later involvement with artificial intelligence and, hence, as a crucial twenty-first century skill. The introduction of CT in primary mathematics curricula puts many demands on teachers, and their understanding of CT in mathematics is key to its successful introduction. Inspired by an information ecology perspective, we investigate how four primary school teachers understand CT in mathematics and how they go ahead to include CT in their mathematics teaching practice. Through observations and interviews, we find promising starting points for including CT, related to pattern recognition, problem solving and the use of programming activities. Our findings indicate that teachers’ lack of knowledge affects CT adoption in two ways: during its inclusion in the existing mathematics curriculum and as a new element focussed on programming and coding, leaving mathematics in the background. For the inclusion to be fruitful, we suggest there is a need to help teachers understand how CT can be used productively in mathematics and vice versa.
- ZeitschriftenartikelIn Memoriam Pamela McCorduck(KI - Künstliche Intelligenz: Vol. 36, No. 1, 2022) Piel, Helen; Seising, Rudolf