This line supports Türkiye’s work on Turkish language models, including BİLGE, the national large language model developed at the TÜBİTAK BİLGEM Artificial Intelligence Institute, and the multinational Turkic LLM initiative.

Contributions include TR-TEB, a standardized benchmark for Turkish text embeddings that evaluates 45 open-source models (LREC 2026), explainable bad-check risk assessment combining machine learning with large language models (SIU 2026), a metric for the conceptual diversity of text, and multimodal speech emotion and text sentiment analysis (UBMK 2023).