Paderborn University

This article presents a multi-level annotation approach for argumentative learner texts that was developed as part of an interdisciplinary DFG project. The project aims at the automated generation of individualized, development-promoting and learner-sensitive feedback on argumentative student texts and is situated in the field of AI-supported text production. To generate automated feedback, the first step was to manually annotate an extensive text corpus consisting of 1,320 argumentative texts written by fifth and ninth graders. This then formed the basis for the development of corresponding computational linguistic procedures. The article focuses on the special features as well as the challenges that arose in connection with the annotation of learner texts and the generation of learner-sensitive feedback. The article is structured as follows: First, the relevant computational linguistics and language didactics research findings and digital support systems for argumentative writing are outlined. In the main part, the procedure of multi-level annotation is explained in detail. Due to the methodological approach, above-average inter-annotator agreement was achieved resulting in the multi-level approach implemented being adaptable for further corpus-based studies. Finally, the results are interpreted and discussed.

To achieve carbon neutrality in urban housing, the buildings’ retrofit and modernisation of their energy systems are needed, implementing the new materials for building envelops, intelligent strategies for heating system control and transferring to renewable sources of energy. All the renovation actions can be implemented in sequence. The present paper proposes an approach based on the mathematical modelling of a building’s energy system efficiency, which estimates the best sequence of the renovation actions. The method considers three groups of renovation actions, containing the (1) thermal modernisation of the building’s envelope, (2) the implementation of smart control systems and (3) the utilisation of renewable energy sources for heat production. To evaluate the effectiveness of individual and combined renovation actions of various technology groups, an integrated relative energy-ecological and economic indicator was introduced. This indicator serves as a basis for determining optimal technical solutions for practical implementation. The proposed approach is applied to the case study of a university building located in Kharkiv, Ukraine. The analysis shows that the application of a smart control system increased the annual relative energy efficiency by 21.5%, and the implementation of the air heat pump increased the integrated relative energy-ecological and economic performance indicator by 47%. Graphical abstract

While existing studies have extensively explored various facets of international students’ experiences, a gap remains in understanding the connection between their social capital accumulation and information-seeking practices that shape the incoming students’ extended transition process. Successful information-seeking is not solely related to academic outcome; it is also influenced by the social capital resources available within and beyond familiar cultural groups. This article sheds light on how international students’ social capital accumulation shapes their information searches across various online and offline social networks. A total of ten international students were interviewed at the beginning and the end of their first semester in Germany during the 2022–23 academic year. The findings demonstrate that the incoming students’ information-seeking strategies and social capital accumulation change over time. The students rely on their previously established social connections offline and online when preparing for their departure. Upon arrival in Germany, the students continue to search for information within their linguistic and cultural familiar groups, but their information-seeking process shifts to local in-person contexts. During the semester, lecturers and fellow students become useful information sources at the university while the students develop confidence in their foreign language skills.

Processability Theory (PT) is a psycholinguistic theory of second language acquisition. The theory builds on the fundamental assumption that learners can acquire only those linguistic forms and functions which they can process. Therefore, PT is based on the architecture of the human language processor. PT is implemented in a theory of grammar that is compatible with the basic design of the language processor. This Element gives a concise introduction to the psycholinguistic core of PT - showing that PT offers an explanation of language development and variation based on processing constraints that are specified for typologically different languages and that apply to first and second language acquisition, albeit in different ways. Processing constraints also delineate transfer from the first language and the effect of formal intervention. This Element also covers the main branches of research in the PT framework and provides an introduction to the methodology used in PT-based research.

The integration of users and experts in machine learning is a widely studied topic in artificial intelligence literature. Similarly, human-computer interaction research extensively explores the factors that influence the acceptance of AI as a decision support system. In this experimental study, we investigate users’ preferences regarding the integration of experts in the development of such systems and how this affects their reliance on these systems. Specifically, we focus on the process of feature selection—an element that is gaining importance due to the growing demand for transparency in machine learning models. We differentiate between three feature selection methods: algorithm-based, expert-based, and a combined approach. In the first treatment, we analyze users’ preferences for these methods. In the second treatment, we randomly assign users to one of the three methods and analyze whether the method affects advice reliance. Users prefer the combined method, followed by the expert-based and algorithm-based methods. However, the users in the second treatment rely equally on all methods. Thus, we find a remarkable difference between stated preferences and actual usage, revealing a significant attitude-behavior-gap. Moreover, allowing the users to choose their preferred method had no effect, and the preferences and the extent of reliance were domain-specific. The findings underscore the importance of understanding cognitive processes in AI-supported decisions and the need for behavioral experiments in human-AI interactions.

A Neumann-type initial-boundary value problem for $\begin{aligned} \left\{ \begin{array}{l} u_{tt} = \nabla \cdot (\gamma (\Theta ) \nabla u_t) + a \nabla \cdot (\gamma (\Theta ) \nabla u) + \nabla \cdot f(\Theta ), \\ \Theta _t = D\Delta \Theta + \Gamma (\Theta ) |\nabla u_t|^2 + F(\Theta )\cdot \nabla u_t, \end{array} \right. \end{aligned}$is considered in a smoothly bounded domain $\Omega \subset \mathbb {R}^n$, $n\ge 1$. In the case when n=1, $\gamma \equiv \Gamma$ and $f\equiv F$, this system coincides with the standard model for heat generation in a viscoelastic material of Kelvin-Voigt type, well-understood in situations in which $\gamma =const$. Covering scenarios in which all key ingredients $\gamma ,\Gamma ,f$ and F may depend on the temperature $\Theta$ here, for initial data which merely satisfy $u_0\in W^{1,p+2}(\Omega )$, $u_{0t}\in W^{1,p}(\Omega )$ and $\Theta _0\in W^{1,p}(\Omega )$ with some $p\ge 2$ such that $p>n$, a result on local-in-time existence and uniqueness is derived in a natural framework of weak solvability.

The execution of incompatible actions imposes costs on action planning, commonly known as response-response incompatibility-costs. This phenomenon is also evident in sports: A basketball player who performs a pass in one direction whilst orienting the head into the contrary direction (pass with head fake) needs more time to initiate the action as if pass direction and head orientation are the same (pass without head fake). The aim of this study was twofold: First, we present a re-analysis of the data from Böer et al. (Psychological Research 88:523–524, 2024) using mixture effect modelling (Miller, Behavior Research Methods 38:92–106, 2006) explore if fake-production costs manifest continuously (uniform effect) in all participants or if some participants show fake-production costs occasionally but substantially (mixed effect). Second, we collected data of a control group which was analysed with the previous data of the practice group and fitted initiation times (ITs) to an ex-Gaussian distribution. The analysis of mixture effects revealed that most participants exhibited a uniform effect when they didn’t have time to mentally prepare the movement. This pattern was not changed by practice, suggesting fake-production costs can’t be overcome by practice alone without mental preparation time. The analysis of mean ITs revealed improvements in the practice group but not in the control group, independent of the type of pass performed. The distribution analyses complemented these findings as it showed that the improvement in participants’ performance with increasing practice can mainly be attributed to a reduction of the exponential part of the distribution (parameter tau).

Autonomous underwater vehicles (AUVs) present several challenges due to the complex and simultaneous interplay of various factors, including but not limited to unmodeled dynamics , highly nonlinear behavior, intercouplings, communication delays, and environmental disturbances. In particular, environmental disturbances degrade trajectory tracking performance for model-based controllers, e.g., model predictive control (MPC) algorithms. Data-driven methods such as the Gaussian process (GP) are effective at learning disturbances in real time; however, the underlying offline hyperparameter tuning process limits their overall effectiveness. To overcome this limitation, we propose a novel dynamic forgetting GP (DF-GP) methodology that compensates for operational disturbances, thus circumventing the need for hyperparameter retuning. In essence, the proposed method optimally combines the predictions of individual GPs-designed with handcrafted forgetting factors, rendering precise disturbance estimation of varying timescales. What is more, the predicted disturbances update the model parameters in MPC, facilitating a learning-based control framework that ensures accurate tracking performance in different underwater scenarios. Rigorous simulation and real-world experiments demonstrate the efficiency and efficacy of the proposed framework. The results show a 25% improvement in disturbance estimation and tracking performance, demonstrating that the proposed framework outperforms its direct competitors.

Coherence refers to correlations between field vibrations at two separate points in degrees of freedom such as space, time, and polarization. In the context of space, coherence theory has been formulated between two transverse positions which can be described either in the Cartesian coordinates or in the cylindrical coordinates. When expressed in cylindrical coordinates, spatial coherence is described in terms of azimuthal and radial coordinates. The description of spatial coherence in radial degree of freedom has been formulated only recently in JOSA A 40, 411 (2023)10.1364/JOSAA.474724. In the present article, we demonstrate an efficient experimental technique for measuring radial coherence, and we report measurement of radial coherence of two different types of radially partially coherent optical fields.

Random fluctuations in somatosensory signals affect the ability of effectively coordinating multimodal information pertaining to the postural state during movement. Therefore, this study aimed to investigate the impact of a compliant surface on cortico-cortical causal information flow during multi-joint compound movements. Fifteen healthy adults (7 female / 8 male, 25.9 ± 4.0 years) performed 5 × 20 repetitions of bodyweight squats on firm and compliant surface. Motor behavior was quantified by center of pressure (CoP) displacements, hip movement and the root mean square of the rectus femoris activity. Using source space analysis, renormalized partial directed coherence (rPDC) computed subject-level multivariate effective brain connectivity of sensorimotor nodes. Bootstrap statistics revealed significantly decreased medio-lateral CoP displacement (p < 0.001), significantly increased velocity of medio-lateral hip motion (p < 0.001) as well as significantly lower rectus femoris activity (p < 0.01) in the compliant surface condition. On the cortical level, rPDC showed significantly modulated information flow in theta and beta frequencies for fronto-parietal edges (p < 0.01) only during the concentric phase of the movement. The compliant surface led to increased difficulties controlling hip but not center of pressure motion in the medio-lateral plane. Moreover, a decreased activation of the prime movers accompanied by modulations of effective brain connectivity among fronto-central nodes may point to altered demands on sensorimotor information processing in presence of sensory noise when performing bodyweight squats on compliant surface. Further studies are needed to evaluate a potential benefit for athletic and clinical populations.