arXiv subjects

Ahmet Tuğrul Bayrak

Publications and source records attributed to Ahmet Tuğrul Bayrak.

4 recordsLinked to original sources

arXiv2026

Comparing Chunking and Embedding Strategies for Turkish RAG Systems

Retrieval-Augmented Generation conditions a language model on chunks retrieved from a document collection. Its accuracy is therefore limited by the chunking and embedding stages that determine what can be retrieved. We compare Turkish document question answering across three chunking strategies (fixed-length, semantic, and layout-aware Docling), five embedding models, and two LLMs, over three documents with contrasting layouts. Every configuration answers the same question set, which allows component effects to be separated by paired testing rather than inferred from separate benchmarks. The fully crossed design yields 9{,}000 graded question-answer evaluations, each scored by an independent judge model, and component comparisons are tested by paired McNemar tests under Holm correction. The three leading embedding models are statistically indistinguishable, so language specialization yields no measurable retrieval advantage. The faster LLM is not the more accurate one. The preferred configuration depends on content type, since layout-aware chunking helps table-heavy documents far more than text-heavy ones.

cs.CL↗

arXiv2026

Real-TurnTurk: A Multimodal Turkish Corpus for Turn-Taking Prediction

Turn-taking is a basic organizational feature of human conversation and remains difficult to model in natural, synchronous dialog systems. While existing research has explored multimodal approaches and large language models for turn-ending prediction, there is a lack of naturalistic conversational corpora specifically addressing turn-taking dynamics in Turkish. This study introduces a multimodal Turkish conversational dataset of unscripted dyadic interactions, comprising synchronized front-facing video, per-speaker audio channels that allow overlapping speech to be attributed to individual speakers, and time-aligned transcriptions. Turn-taking prediction is formulated as a binary classification problem, and a Genetic Algorithm (GA) is employed to optimize interpretable decision rules derived from visual, acoustic, and linguistic features. A hybrid AND-OR rule representation is adopted in the proposed framework to represent the alternative cue combinations that precede a turn transition.

cs.CL↗

arXiv2026

Syn-TurnTurk: A Synthetic Dataset for Turn-Taking Prediction in Turkish Dialogues

Managing natural dialogue timing is a significant challenge for voice-based chatbots. Most current systems usually rely on simple silence detection, which often fails because human speech patterns involve irregular pauses. This causes bots to interrupt users, breaking the conversational flow. This problem is even more severe for languages like Turkish, which lack high-quality datasets for turn-taking prediction. This paper introduces Syn-TurnTurk, a synthetic Turkish dialogue dataset generated using various Qwen Large Language Models (LLMs) to mirror real-life verbal exchanges, including overlaps and strategic silences. We evaluated the dataset using several traditional and deep learning architectures. The results show that advanced models, particularly BI-LSTM and Ensemble (LR+RF) methods, achieve high accuracy (0.839) and AUC scores (0.910). These findings demonstrate that our synthetic dataset can have a positive affect for models understand linguistic cues, allowing for more natural human-machine interaction in Turkish.

cs.CL↗

arXiv2022

Using Person Embedding to Enrich Features and Data Augmentation for Classification

Today, machine learning is applied in almost any field. In machine learning, where there are numerous methods, classification is one of the most basic and crucial ones. Various problems can be solved by classification. The feature selection for model setup is extremely important, and producing new features via feature engineering also has a vital place in the success of the model. In our study, fraud detection classification models are built on a labeled and imbalanced dataset as a case-study. Although it is a natural language processing method, a customer space has been created with word embedding, which has been used in different areas, especially for recommender systems. The customer vectors in the created space are fed to the classification model as a feature. Moreover, to increase the number of positive labels, rows with similar characteristics are re-labeled as positive by using customer similarity determined by embedding. The model in which embedding methods are included in the classification, which provides a better representation of customers, has been compared with other models. Considering the results, it is observed that the customer embedding method had a positive effect on the success of the classification models.

cs.LG↗