TURSpider Veri K mesinde Temsilcilerin Kari simi Tabanli Text-to-SQL ali smasi

Ali Bugra Kanburoglu; F. Boray Tek

doi:10.1109/SIU66497.2025.11112412

TURSpider Veri K mesinde Temsilcilerin Kari simi Tabanli Text-to-SQL ali smasi

Translated title of the contribution: Mixture-of-Agents based Text-to-SQL Study on TURSpider Dataset

Ali Bugra Kanburoglu^*
, F. Boray Tek

^*Corresponding author for this work

Department of Artificial Intelligence and Data Engineering

Iik Üniversitesi

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

1 Citation (Scopus)

Abstract

This study focuses on experiments conducted on the TURSpider dataset, developed for the Turkish Text-to-SQL task. TURSpider is a large-scale Turkish dataset containing SQL queries of varying difficulty levels and serves as a valuable resource for research in this field. The study investigates the effectiveness of the feedback-driven Mixture-of-Agents (MoAF) approach on this task. In the MoAF structure, multiple large language models (LLMs) collaborate to improve SQL generation performance. In this setup, agent collaboration enables models to learn from each other and correct errors through feedback mechanisms. According to the experimental results, the MoAF approach achieved an execution accuracy of 60.63%, achieving the highest performance reported on the TURSpider dataset.

Translated title of the contribution	Mixture-of-Agents based Text-to-SQL Study on TURSpider Dataset
Original language	Turkish
Title of host publication	33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025 - Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)	9798331566555
DOIs	https://doi.org/10.1109/SIU66497.2025.11112412
Publication status	Published - 2025
Event	33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025 - Istanbul, Turkey Duration: 25 Jun 2025 → 28 Jun 2025

Publication series

Name	33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025 - Proceedings

Conference

Conference	33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025
Country/Territory	Turkey
City	Istanbul
Period	25/06/25 → 28/06/25

Bibliographical note

Publisher Copyright:
© 2025 IEEE.

Access to Document

10.1109/SIU66497.2025.11112412

Cite this

Kanburoglu, A. B., & Tek, F. B. (2025). TURSpider Veri K mesinde Temsilcilerin Kari simi Tabanli Text-to-SQL ali smasi. In 33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025 - Proceedings (33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025 - Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/SIU66497.2025.11112412

Kanburoglu, Ali Bugra ; Tek, F. Boray. / TURSpider Veri K mesinde Temsilcilerin Kari simi Tabanli Text-to-SQL ali smasi. 33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2025. (33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025 - Proceedings).

@inproceedings{875a23a5b9b14bb081065aa9febe0c96,

title = "TURSpider Veri K mesinde Temsilcilerin Kari simi Tabanli Text-to-SQL ali smasi",

abstract = "This study focuses on experiments conducted on the TURSpider dataset, developed for the Turkish Text-to-SQL task. TURSpider is a large-scale Turkish dataset containing SQL queries of varying difficulty levels and serves as a valuable resource for research in this field. The study investigates the effectiveness of the feedback-driven Mixture-of-Agents (MoAF) approach on this task. In the MoAF structure, multiple large language models (LLMs) collaborate to improve SQL generation performance. In this setup, agent collaboration enables models to learn from each other and correct errors through feedback mechanisms. According to the experimental results, the MoAF approach achieved an execution accuracy of 60.63\%, achieving the highest performance reported on the TURSpider dataset.",

keywords = "Chain-of-Feedback (CoF), Mixture-of-Agents (MoA), TURSpider, Text-to-SQL, Turkish",

author = "Kanburoglu, \{Ali Bugra\} and Tek, \{F. Boray\}",

note = "Publisher Copyright: {\textcopyright} 2025 IEEE.; 33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025 ; Conference date: 25-06-2025 Through 28-06-2025",

year = "2025",

doi = "10.1109/SIU66497.2025.11112412",

language = "T{\"u}rk{\c c}e",

series = "33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025 - Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

booktitle = "33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025 - Proceedings",

address = "United States",

}

Kanburoglu, AB & Tek, FB 2025, TURSpider Veri K mesinde Temsilcilerin Kari simi Tabanli Text-to-SQL ali smasi. in 33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025 - Proceedings. 33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025 - Proceedings, Institute of Electrical and Electronics Engineers Inc., 33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025, Istanbul, Turkey, 25/06/25. https://doi.org/10.1109/SIU66497.2025.11112412

TURSpider Veri K mesinde Temsilcilerin Kari simi Tabanli Text-to-SQL ali smasi. / Kanburoglu, Ali Bugra; Tek, F. Boray.
33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2025. (33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025 - Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - TURSpider Veri K mesinde Temsilcilerin Kari simi Tabanli Text-to-SQL ali smasi

AU - Kanburoglu, Ali Bugra

AU - Tek, F. Boray

PY - 2025

Y1 - 2025

N2 - This study focuses on experiments conducted on the TURSpider dataset, developed for the Turkish Text-to-SQL task. TURSpider is a large-scale Turkish dataset containing SQL queries of varying difficulty levels and serves as a valuable resource for research in this field. The study investigates the effectiveness of the feedback-driven Mixture-of-Agents (MoAF) approach on this task. In the MoAF structure, multiple large language models (LLMs) collaborate to improve SQL generation performance. In this setup, agent collaboration enables models to learn from each other and correct errors through feedback mechanisms. According to the experimental results, the MoAF approach achieved an execution accuracy of 60.63%, achieving the highest performance reported on the TURSpider dataset.

AB - This study focuses on experiments conducted on the TURSpider dataset, developed for the Turkish Text-to-SQL task. TURSpider is a large-scale Turkish dataset containing SQL queries of varying difficulty levels and serves as a valuable resource for research in this field. The study investigates the effectiveness of the feedback-driven Mixture-of-Agents (MoAF) approach on this task. In the MoAF structure, multiple large language models (LLMs) collaborate to improve SQL generation performance. In this setup, agent collaboration enables models to learn from each other and correct errors through feedback mechanisms. According to the experimental results, the MoAF approach achieved an execution accuracy of 60.63%, achieving the highest performance reported on the TURSpider dataset.

KW - Chain-of-Feedback (CoF)

KW - Mixture-of-Agents (MoA)

KW - TURSpider

KW - Text-to-SQL

KW - Turkish

UR - https://www.scopus.com/pages/publications/105015478722

U2 - 10.1109/SIU66497.2025.11112412

DO - 10.1109/SIU66497.2025.11112412

M3 - Konferans katkısı

AN - SCOPUS:105015478722

T3 - 33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025 - Proceedings

BT - 33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025

Y2 - 25 June 2025 through 28 June 2025

ER -

Kanburoglu AB, Tek FB. TURSpider Veri K mesinde Temsilcilerin Kari simi Tabanli Text-to-SQL ali smasi. In 33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2025. (33rd IEEE Conference on Signal Processing and Communications Applications, SIU 2025 - Proceedings). doi: 10.1109/SIU66497.2025.11112412

TURSpider Veri K mesinde Temsilcilerin Kari simi Tabanli Text-to-SQL ali smasi

Abstract

Publication series

Conference

Bibliographical note

Access to Document

Other files and links

Fingerprint

Cite this