Kurdish Natural Language Understanding (NLU) and Text Classification investigates deep learning and contextual representation models designed to understand semantics, categorize documents, and extract information from Kurdish texts.
Research Objectives
The research explores adapting pretrained language models and transformer architectures to Kurdish linguistic structures, evaluating semantic comprehension across diverse domains and genres.
Key Areas
- Multi-class document and topic classification
- Named Entity Recognition (NER) and semantic role labeling
- Intent recognition and contextual semantic understanding
- Evaluating fine-tuned transformer models for Kurdish
Applications
This work provides foundational tools for automated content indexing, digital archiving, legal and academic document organization, and smart information retrieval in Kurdish.