A5063061204- Education Practices and Challenges
- Natural Language Processing Techniques
- Topic Modeling
- Educational Methods and Analysis
- Educational Leadership and Administration
- Teacher Professional Development and Motivation
- Online and Blended Learning
- Music Education and Analysis
- Speech Recognition and Synthesis
- Innovative Teaching and Learning Methods
- Virtual Reality Applications and Impacts
- Text Readability and Simplification
- Speech and dialogue systems
- Online Learning and Analytics
- Education and Technology Integration
- Multimodal Machine Learning Applications
- Problem Solving Skills Development
- Mobile Learning in Education
- Public Administration and Governance
- Technology-Enhanced Education Studies
- Augmented Reality Applications
- Mobile and Web Applications
- Education and experiences of immigrants and refugees
- Values and Moral Education
- Domain Adaptation and Few-Shot Learning
Bartin University
2017-2024
Amasya Üniversitesi
2009-2023
University of Groningen
2019-2022
Kayseri Eğitim ve Araştırma Hastanesi
2022
Mi̇lli̇ Eği̇ti̇m Bakanliği
2022
UCSI University
2022
University of Edinburgh
2022
IT University of Copenhagen
2021
The Microsoft Research - University of Trento Centre for Computational and Systems Biology
2021
University of Trento
2021
Recent advances in multilingual dependency parsing have brought the idea of a truly universal parser closer to reality. However, cross-language interference and restrained model capacity remain major obstacles. To address this, we propose novel task adaptation approach based on contextual parameter generation adapter modules. This enables learn adapters via language embeddings while sharing parameters across languages. It also allows for an easy but effective integration existing linguistic...
Rob van der Goot, Ahmet Üstün, Alan Ramponi, Ibrahim Sharaf, Barbara Plank. Proceedings of the 16th Conference European Chapter Association for Computational Linguistics: System Demonstrations. 2021.
This technical report introduces Aya 23, a family of multilingual language models. 23 builds on the recent release model (\"Ust\"un et al., 2024), focusing pairing highly performant pre-trained with recently released collection (Singh 2024). The result is powerful large serving languages, expanding state-of-art modeling capabilities to approximately half world's population. covered 101 languages whereas an experiment in depth vs breadth, exploring impact allocating more capacity fewer that...
This mixed-method pretest/post-test experimental study examined the effect of learning analytics (LA)-based interventions on students' academic achievement and self-regulatory skills, explored perceptions such in flipped classrooms (FC). Sixty-two college students from an introductory computer course were randomly assigned to or control group, with 31 participants each condition. In following 10 weeks, LA-based personalized interventions, including both visual feedback written...
Blended learning (BL) has been increasingly implemented in higher education, while more research is needed to investigate the role of metacognitive awareness, reflective thinking, problem solving and community inquiry on academic self-efficacy BL. This correlational study collected data from 217 undergraduate students an introductory computer course, using five well-established instruments. The result showed that there was a strong positive relationship between thinking skills, also moderate...
Rob van der Goot, Ibrahim Sharaf, Aizhan Imankulova, Ahmet Üstün, Marija Stepanović, Alan Ramponi, Siti Oryza Khairunnisa, Mamoru Komachi, Barbara Plank. Proceedings of the 2021 Conference North American Chapter Association for Computational Linguistics: Human Language Technologies. 2021.
We consider the problem of multilingual unsupervised machine translation, translating to and from languages that only have monolingual data by using auxiliary parallel language pairs. For this standard procedure so far leverage is _back-translation_, which computationally costly hard tune. In paper we propose instead use _denoising adapters_, adapter layers with a denoising objective, on top pre-trained mBART-50. addition modularity flexibility such an approach show resulting translations...
Değişen ve gelişen dünyaya uyum sağlayabilmek için insanların da aynı değişim gelişime ayak uydurabilecek becerilere sahip olması beklenmektedir. Eğitim alanında hem öğrenenlerin de öğretmenlerin yeni büyük önem taşımaktadır. Bu kapsamda eğitimde öğrencilerin ilgi motivasyonu artıracak yöntem teknikler kullanılmaktadır. yöntemlerden biri mobil teknolojiler kullanılarak öğrenme destekli öğretim gerçekleştirmektir. bağlamda sürecini destekleyici uygulamaları ön plana çıkmaktadır. çalışmada...
Within higher education, the use of blended learning (BL) is exponentially increased in 21st century, which poses a number challenges designing process BL for educators. The aim this designbased research (DBR) study was to assist an inexperienced educator teaching and course education convert face-to-face (F2F) into course. During design delivery course, what appropriate practices were needed achieve creating efficient effective determined educator's reflections on first experience...
The Mixture of Experts (MoE) is a widely known neural architecture where an ensemble specialized sub-models optimizes overall performance with constant computational cost. However, conventional MoEs pose challenges at scale due to the need store all experts in memory. In this paper, we push MoE limit. We propose extremely parameter-efficient by uniquely combining lightweight experts.Our outperforms standard fine-tuning (PEFT) methods and on par full only updating -- less than 1% 11B...
In this paper, we investigate the effects of using subword information in representation learning. We argue that syntactic units quality word representations positively. introduce a morpheme-based model and compare it against to word-based, character-based, character n-gram level models. Our takes list candidate segmentations learns based on different are weighted by an attention mechanism. performed experiments Turkish as morphologically rich language English with comparably poorer...
The use of flipped learning (FL) is increasing in higher education with enormous educational benefits. But limited research has investigated the relationships students' transactional distance (TD) perceptions, engagement, and course satisfaction FL. This correlational study was conducted to explore these important factors their 198 undergraduate students participated this 14-week-long study, including 105 females 93 males. Three well-established instruments were employed collect data. path...
Recent breakthroughs in large language models (LLMs) have centered around a handful of data-rich languages. What does it take to broaden access beyond first-class citizen languages? Our work introduces Aya, massively multilingual generative model that follows instructions 101 languages which over 50% are considered as lower-resourced. Aya outperforms mT0 and BLOOMZ on the majority tasks while covering double number We introduce extensive new evaluation suites state-of-art for eval across 99...