SynTran-fa: Generating Comprehensive Answers for Farsi QA Pairs via Syntactic Transformation

• DOI: 10.20944/preprints202410.1684.v1
• Publication Date: 2024-10-23T01:56:03Z
• Authors: Farhan Farsi; Sadra Sabouri; Kian Kashfipour; Soroush Gooran; Hossein Sameti; Ehsaneddin Asgari

Generating coherent and comprehensive responses remains a significant challenge Question-Answering (QA) systems when working with short answers especially for low-resourced languages like Farsi. We present novel approach to expand these into complete, fluent responses, addressing the critical issue of limited Farsi resources models. Our methodology employs two-stage process: first, we develop dataset using rule-based techniques on text, followed by BERT-based ranking system ensure fluency comprehensibility. The resulting model demonstrates strong compatibility existing QA systems, particularly those based knowledge graphs. Notably, our exhibits enhanced performance integrated large language models Chain-of-Thought (CoT) prompting, leveraging detailed explanations rather than single-word answers. significantly improves response quality coherence compared baseline systems. release support further research in QA\footnote[1]{\href{https://huggingface.co/datasets/SLPL/syntran-fa}{https://huggingface.co/datasets/SLPL/syntran-fa}}.