Masters

Permanent URI for this collection

https://repository.sdu.edu.kz/handle/123456789/1256

Browse

Browsing Masters by Subject "AI in education"

Now showing 1 - 2 of 2
  • Thumbnail Image
    ItemOpen Access
    A Comparative Study of AI-assisted Assessment and Teacher-Assessment in an EFL Writing Course
    (SDU University, 2025) Fazilova A.
    With the rapid development of artificial intelligence in education, automated writing evaluation (AWE) tools such as ChatGPT are becoming increasingly popular for providing feedback on students' written works. This study explores the perceptions of first-year EFL students about AI-assisted assessment compared to teacher assessment in an EFL writing course at a private Kazakhstani university. Over the period of one semester, 33 participants have written 4 essays and received both the teacher and ChatGPT feedback on each essay. Their perceptions have been compared and analyzed with a quantitative research design with elements of qualitative analysis. A questionnaire with Likert-scale closed-ended and open-ended questions was used. The results revealed that while students recognize the importance of AI-assisted assessment for surface-level corrections (grammar, vocabulary, structure), the majority of students prefer teacher feedback for its clarity, personalized support, and depth. Additionally, most students viewed the ideal approach as a combination of two types of feedback: AI for quick technical feedback and teachers for more complex aspects like structure, argumentation, and tone. The study concludes that although AWE tools have potential as supplementary support in EFL writing instruction, they cannot replace the human connection and pedagogical insight offered by teachers. Implications for integrating AI tools into classroom practice and teacher training are also discussed, along with recommendations for future research in this evolving field.
  • Thumbnail Image
    ItemOpen Access
    Detecting social conflicts in kindergartens using deep learning and computer vision
    (SDU University, 2025) Kengesbay D.
    Early conflict detection in kindergartens plays a significant role in ensuring a harmonious learning atmosphere and in promoting the social growth of young children. While most previous works have only addressed conflict detection through adults, in this paper, we specifically address conflict detection in kindergartens using deep learning, utilizing both spatial and temporal information to improve performance. The application of deep learning and computer vision in automatically detecting and analyzing early conflicts among young children is discussed in this paper. Using video footage, we leverage state-of-theart RNNs and 3D CNNs for high-accuracy detection of conflict instances. Crucial visual cues—facial expressions, gestures, poses, vocal tone, and movement—are examined for the extraction of tension or aggression signs. The model is evaluated on real kindergarten video data, with promising conflict detection and classification results. The findings indicate the potential of AI-supported tools in assisting teachers in class management, child behavior monitoring, early intervention mechanisms, and the fostering of a good social environment