Middle East Technical University Research Information System
Not a member yet
104897 research outputs found
Sort by
The use of artificially intelligent chatbots in English language learning: A systematic meta-synthesis study of articles published between 2010 and 2024
No full textIn this qualitative systematic meta-synthesis study, 57 studies from the international literature published between 2010 and 2024 on the use of voice-based artificially intelligent chatbots in English language learning were analyzed. The present study aimed to explore the most recent studies on this topic by investigating the theoretical frameworks, methodological and technological properties, user reports of chatbot usage experience, and pedagogical implementations. It sought to identify research and implementation trends for voice-based chatbots via qualitative data analysis methods. Based on the reviewed studies, this paper presents data-based pedagogical implications that align with the latest voice-based AI chatbot research trends
The impact of COVID-19 on the social and cultural integration of international students: a literature review
No full textThis systematic literature review summarises the state-of-the-art evidence on the impact of COVID-19 on the integration of international students in their host countries and institutions. Conducted between January and May 2022, it analyses the responses to COVID-19 of the key actors involved in international student mobility: national/regional authorities, higher education institutions, and students. Findings reveal that governmental action and institutional measures were decisive in shaping international students’ integration experiences. Regarding governmental action, criticism of the policies adopted by Australia and the USA in relation to immigration and/or support stand out, in contrast to policies adopted by the Canadian authorities. Higher education institutions played an important role in mitigating the negative effects of COVID-19 on international students’ integration. These targeted different needs– material, well-being, and social– through different types of support: logistical and financial support, psychological support, and the provision of platforms for ongoing social interaction and exchange. Most studies, however, focus on the students themselves, the challenges they faced during the pandemic and their coping strategies. Common to international students’ lived experience was (dis)connectedness, with the following themes emerging as obstacles to their social and cultural integration: distress during lockdown periods, disruption of their social life and support networks, mental health issues, discrimination and racialised prejudice, and language barriers. The review concludes by proposing recommendations and by identifying avenues for future research
Low-Level Hardware Requirement Classification Using Large Language Models: Challenges, Insights, and Future Directions for Embedded Control Systems
No full textAutomated Requirements Engineering (RE) activities can streamline development processes, reduce errors, and facilitate informed decision-making, particularly for low-level hardware requirements where modifications are costly. Classification is a widely studied automated RE activity for software requirements. Yet, its applicability remains underexplored due to the lack of structured datasets. This study adapts and evaluates software requirement classification techniques for hardware by extracting low-level requirements from open-source hardware design artifacts of Embedded Control Systems. We evaluate two classification methods: fine-tuning a BERT-based model and zero-shot prompting with a quantized LLM (Qwen2.5). While fine-tuning achieved high accuracy, zero-shot classification with specific prompts outperformed it in overall performance, achieving an average F1-score of up to 90% on the hold-out test set. Our findings suggest that automating downstream RE activities for low-level hardware requirements may not require large, task-specific datasets; however, classification performance can be further improved and can serve as an enabler for advanced tasks
Magnetic and luminescence properties of bioactive glass nanoparticles for biomedical applications
No full textThis study explores the synthesis and characterization of superparamagnetic iron oxide nanoparticles (SPIONs) coated with zinc (Zn) and/or europium (Eu) doped bioactive glass. Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy (TEM) confirmed spherical agglomerated morphology and core@shell structure, respectively. High-Resolution TEM (HR-TEM) revealed lattice fringe values consistent with the cubic magnetite phase. Magnetic property assessment showed stable superparamagnetic behavior with slight reductions in saturation magnetization (sigma s) after immersion in simulated body fluid (SBF) solution. Photoluminescence (PL) spectra of Eu-doped samples exhibited red emission, confirming Eu rare earth element incorporation and maintaining luminescence post-immersion in SBF. Upon the interaction with SBF, hydroxyapatite (HA) formation occurred on the nanoparticle surfaces, suggesting the bioactive nature of the nanoparticles. These findings suggest that the synthesized nanoparticles exhibit promising potential for biomedical applications, including imaging, and orthopedics, due to their bioactive, magnetic, and luminescent properties
High-performance bifunctional electrochromic-supercapacitor devices based on indole derivative copolymerized with 3,4-ethylenedioxythiophene
No full textThe integration of energy storage and electrochromic functionalities into a single device holds great promise for smart windows, displays, and energy-efficient systems. This study presents the first bifunctional electrochromic supercapacitor device (ECSCD) based on an indole-7-carboxylic acid (IN7Ca) and 3,4-ethylenedioxythiophene (EDOT) copolymer, synthesized via electropolymerization. Unlike conventional single-material systems, this work leverages a novel asymmetric design (PIN7Ca anode/P(EDOT-co-IN7Ca) cathode) to simultaneously achieve record performance in both optical and energy storage metrics. The electrochromic and supercapacitor properties of the homopolymer (PIN7Ca), copolymer (P(EDOT-co-IN7Ca)), and PEDOT are systematically investigated. The copolymer demonstrates unprecedented optical transmittance, faster switching (1.8 s), and enhanced electrochromic performance due to the incorporation of EDOT. Its improved capacitive behavior and >90 % cycling stability after 2000 cycles arise from the synergistic combination of IN7Ca's redox activity and EDOT's electrochromic stability. An asymmetric ECSCD achieves a record electrochromic contrast (ΔT = 47 %) among solution-processed devices, alongside a high specific capacitance (1.9 mF/cm² at 0.01 mA/cm²) that surpasses recent PEDOT- and PANI-based systems by >30 %. These results highlight copolymerization as an effective strategy for designing multifunctional materials with benchmark performance. This approach establishes new design principles for high-performance ECSCDs, supporting their integration into next-generation smart and sustainable technologies
Class distance weighted cross entropy loss for classification of disease severity
No full textAssessing disease severity with ordinal classes, where each class reflects increasing severity levels, benefits from loss functions designed for this ordinal structure. Traditional categorical loss functions, like Cross-Entropy (CE), often perform suboptimally in these scenarios. To address this, we propose a novel loss function, Class Distance Weighted Cross-Entropy (CDW-CE), which penalizes misclassifications more severely when the predicted and actual classes are farther apart. We evaluated CDW-CE using various deep architectures, comparing its performance against several categorical and ordinal loss functions. To assess the quality of latent representations, we used t-distributed stochastic neighbor embedding (t-SNE) and uniform manifold approximation and projection (UMAP) visualizations, quantified the clustering quality using the Silhouette Score, and compared Class Activation Maps (CAM) generated by models trained with CDW-CE and CE loss. Feedback from domain experts was incorporated to evaluate how well model attention aligns with expert opinion. Our results show that CDW-CE consistently improves performance in ordinal image classification tasks. It achieves higher Silhouette Scores, indicating better class discrimination capability, and its CAM visualizations show a stronger focus on clinically significant regions, as validated by domain experts. Receiver operator characteristics (ROC) curves and the area under the curve (AUC) scores highlight that CDW-CE outperforms other loss functions, including prominent ordinal loss functions from the literature
Desulfurization of Coal with Oak Ash Extraction
No full textDesulfurization is an essential method to eliminate the harmful environmental effects of coal. Various physical, chemical and biological desulfurization methods are available in the literature. In this study about with the desulfurization of coal by a basic extraction liquid of oak ash. Burdur, Ermenek and Tuncbilek coals were selected due to their high sulfur content. Before and after the desulfurization process; total sulfur, pyritic sulfur, carbon content and calorific value were measured. The highest removal efficiencies of total sulfur and pyritic sulfur for Burdur coal obtained as 41 and 67%, for Tuncbilek coal obtained as 41 and 35%, and for Ermenek coal obtained as 47 and 41%, respectively. Furthermore, the carbon and calorific value of these coals were analyzed and it seen that these values were not affected negatively. These findings confirmed that the desulfurization with oak ash of Burdur, Tuncbilek and Ermenek coals could help as a reference for further optimizing the coal desulfurization process
Uncovering Parental Ethnotheories in Türkiye: Parental Beliefs and Practices Linkage
No full textParental ethnotheories delineate culturally shared beliefs about the nature of children and normative parenting in a particular cultural niche. Using a sequential mixed-methods design, we assessed parental ethnotheories in a non-White, educated, industrialized, rich, and developed cultural context of Türkiye and developed a parental beliefs scale (PBS) with a culturally informed emic approach in two studies. Study 1 relied on semistructured interviews with 125 Turkish parents (79 mothers, 46 fathers) to better understand parents’ beliefs on the child’s nature and proper parenting with particular attention to the key demographic characteristics reflecting intracultural diversity. This qualitative inquiry informed the generation of items for a PBS about the nature of children and parenting. In Study 2, we investigated the factor structure, measurement invariance, and the predictive power of the PBS on parenting behaviors with a nationally representative sample of 1,397 parents (796 mothers, 601 fathers) of children aged 3–17 years. Factor analysis revealed three factors representing constraining beliefs, autonomy-enabling beliefs, and beliefs in the malleability of the child. Structural and measurement invariance analyses partially supported the equivalence of the three-factor structure across parent and child gender and child age groups. Regression analyses indicated that constraining beliefs strongly and positively predicted psychological control and punitive behaviors. Autonomy-enabling beliefs predicted positive parenting, while malleability beliefs primarily predicted sociocultural control. Parent education and socioeconomic status moderated the effects of parental beliefs on parenting behaviors. The results were discussed based on parents’ gender and socioeconomic status within a developing country, exemplifying a culturally informed assessment approach for the majority world