Integrating Question Answering and Text-to-SQL in Portuguese

Marcos Menon José; Marcelo Archanjo José; Denis Deratani Mauá; Fábio Gagliardi Cozman

doi:10.1007/978-3-030-98305-5_26

Integrating Question Answering and Text-to-SQL in Portuguese

Marcos Menon José, Marcelo Archanjo José, Denis Deratani Mauá, Fábio Gagliardi Cozman

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

Abstract

Deep learning transformers have drastically improved systems that automatically answer questions in natural language. However, different questions demand different answering techniques; here we propose, build and validate an architecture that integrates different modules to answer two distinct kinds of queries. Our architecture takes a free-form natural language text and classifies it to send it either to a Neural Question Answering Reasoner or a Natural Language parser to SQL. We implemented a complete system for the Portuguese language, using some of the main tools available for the language and translating training and testing datasets. Experiments show that our system selects the appropriate answering method with high accuracy (over 99%), thus validating a modular question answering strategy.

Original language	English
Title of host publication	Computational Processing of the Portuguese Language - 15th International Conference, PROPOR 2022, Proceedings
Editors	Vládia Pinheiro, Pablo Gamallo, Raquel Amaro, Carolina Scarton, Fernando Batista, Diego Silva, Catarina Magro, Hugo Pinto
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	278-287
Number of pages	10
ISBN (Print)	9783030983048
DOIs	https://doi.org/10.1007/978-3-030-98305-5_26
State	Published - 2022
Externally published	Yes
Event	15th International Conference on the Computational Processing of Portuguese, PROPOR 2022 - Fortaleza, Brazil Duration: 21 Mar 2022 → 23 Mar 2022

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	13208 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	15th International Conference on the Computational Processing of Portuguese, PROPOR 2022
Country/Territory	Brazil
City	Fortaleza
Period	21/03/22 → 23/03/22

Keywords

Natural language interfaces to databases
Natural language processing in portuguese
Question answering
Transformers networks

Access to Document

10.1007/978-3-030-98305-5_26

Cite this

José, M. M., José, M. A., Mauá, D. D., & Cozman, F. G. (2022). Integrating Question Answering and Text-to-SQL in Portuguese. In V. Pinheiro, P. Gamallo, R. Amaro, C. Scarton, F. Batista, D. Silva, C. Magro, & H. Pinto (Eds.), Computational Processing of the Portuguese Language - 15th International Conference, PROPOR 2022, Proceedings (pp. 278-287). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13208 LNAI). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-98305-5_26

José, Marcos Menon ; José, Marcelo Archanjo ; Mauá, Denis Deratani et al. / Integrating Question Answering and Text-to-SQL in Portuguese. Computational Processing of the Portuguese Language - 15th International Conference, PROPOR 2022, Proceedings. editor / Vládia Pinheiro ; Pablo Gamallo ; Raquel Amaro ; Carolina Scarton ; Fernando Batista ; Diego Silva ; Catarina Magro ; Hugo Pinto. Springer Science and Business Media Deutschland GmbH, 2022. pp. 278-287 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{77a36b38720a48b588fa301553afb973,

title = "Integrating Question Answering and Text-to-SQL in Portuguese",

abstract = "Deep learning transformers have drastically improved systems that automatically answer questions in natural language. However, different questions demand different answering techniques; here we propose, build and validate an architecture that integrates different modules to answer two distinct kinds of queries. Our architecture takes a free-form natural language text and classifies it to send it either to a Neural Question Answering Reasoner or a Natural Language parser to SQL. We implemented a complete system for the Portuguese language, using some of the main tools available for the language and translating training and testing datasets. Experiments show that our system selects the appropriate answering method with high accuracy (over 99%), thus validating a modular question answering strategy.",

keywords = "Natural language interfaces to databases, Natural language processing in portuguese, Question answering, Transformers networks",

author = "Jos{\'e}, {Marcos Menon} and Jos{\'e}, {Marcelo Archanjo} and Mau{\'a}, {Denis Deratani} and Cozman, {F{\'a}bio Gagliardi}",

note = "Funding Information: Acknowledgment. This work was partly supported by Ita{\'u} Unibanco S.A. through the Programa de Bolsas Ita{\'u} (PBI) of the Centro de Ci{\^e}ncia de Dados da Universidade de S{\~a}o Paulo (C2D-USP); by the Center for Artificial Intelligence (C4AI) through support from the S{\~a}o Paulo Research Foundation (FAPESP grant #2019/07665-4) and from the IBM Corporation; by CNPq grants no. 312180/2018-7 and 304012/2019-0, and CAPES Finance Code 001. Publisher Copyright: {\textcopyright} 2022, Springer Nature Switzerland AG.; null ; Conference date: 21-03-2022 Through 23-03-2022",

year = "2022",

doi = "10.1007/978-3-030-98305-5_26",

language = "Ingl{\'e}s",

isbn = "9783030983048",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "278--287",

editor = "Vl{\'a}dia Pinheiro and Pablo Gamallo and Raquel Amaro and Carolina Scarton and Fernando Batista and Diego Silva and Catarina Magro and Hugo Pinto",

booktitle = "Computational Processing of the Portuguese Language - 15th International Conference, PROPOR 2022, Proceedings",

address = "Alemania",

}

José, MM, José, MA, Mauá, DD & Cozman, FG 2022, Integrating Question Answering and Text-to-SQL in Portuguese. in V Pinheiro, P Gamallo, R Amaro, C Scarton, F Batista, D Silva, C Magro & H Pinto (eds), Computational Processing of the Portuguese Language - 15th International Conference, PROPOR 2022, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 13208 LNAI, Springer Science and Business Media Deutschland GmbH, pp. 278-287, 15th International Conference on the Computational Processing of Portuguese, PROPOR 2022, Fortaleza, Brazil, 21/03/22. https://doi.org/10.1007/978-3-030-98305-5_26

Integrating Question Answering and Text-to-SQL in Portuguese. / José, Marcos Menon; José, Marcelo Archanjo; Mauá, Denis Deratani et al.

Computational Processing of the Portuguese Language - 15th International Conference, PROPOR 2022, Proceedings. ed. / Vládia Pinheiro; Pablo Gamallo; Raquel Amaro; Carolina Scarton; Fernando Batista; Diego Silva; Catarina Magro; Hugo Pinto. Springer Science and Business Media Deutschland GmbH, 2022. p. 278-287 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13208 LNAI).

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

TY - GEN

T1 - Integrating Question Answering and Text-to-SQL in Portuguese

AU - José, Marcos Menon

AU - José, Marcelo Archanjo

AU - Mauá, Denis Deratani

AU - Cozman, Fábio Gagliardi

N1 - Funding Information: Acknowledgment. This work was partly supported by Itaú Unibanco S.A. through the Programa de Bolsas Itaú (PBI) of the Centro de Ciência de Dados da Universidade de São Paulo (C2D-USP); by the Center for Artificial Intelligence (C4AI) through support from the São Paulo Research Foundation (FAPESP grant #2019/07665-4) and from the IBM Corporation; by CNPq grants no. 312180/2018-7 and 304012/2019-0, and CAPES Finance Code 001. Publisher Copyright: © 2022, Springer Nature Switzerland AG.

PY - 2022

Y1 - 2022

N2 - Deep learning transformers have drastically improved systems that automatically answer questions in natural language. However, different questions demand different answering techniques; here we propose, build and validate an architecture that integrates different modules to answer two distinct kinds of queries. Our architecture takes a free-form natural language text and classifies it to send it either to a Neural Question Answering Reasoner or a Natural Language parser to SQL. We implemented a complete system for the Portuguese language, using some of the main tools available for the language and translating training and testing datasets. Experiments show that our system selects the appropriate answering method with high accuracy (over 99%), thus validating a modular question answering strategy.

AB - Deep learning transformers have drastically improved systems that automatically answer questions in natural language. However, different questions demand different answering techniques; here we propose, build and validate an architecture that integrates different modules to answer two distinct kinds of queries. Our architecture takes a free-form natural language text and classifies it to send it either to a Neural Question Answering Reasoner or a Natural Language parser to SQL. We implemented a complete system for the Portuguese language, using some of the main tools available for the language and translating training and testing datasets. Experiments show that our system selects the appropriate answering method with high accuracy (over 99%), thus validating a modular question answering strategy.

KW - Natural language interfaces to databases

KW - Natural language processing in portuguese

KW - Question answering

KW - Transformers networks

UR - http://www.scopus.com/inward/record.url?scp=85127206125&partnerID=8YFLogxK

UR - https://www.mendeley.com/catalogue/5d90488a-3bf1-39e2-9cb8-905715b926fe/

U2 - 10.1007/978-3-030-98305-5_26

DO - 10.1007/978-3-030-98305-5_26

M3 - Contribución a la conferencia

AN - SCOPUS:85127206125

SN - 9783030983048

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 278

EP - 287

BT - Computational Processing of the Portuguese Language - 15th International Conference, PROPOR 2022, Proceedings

A2 - Pinheiro, Vládia

A2 - Gamallo, Pablo

A2 - Amaro, Raquel

A2 - Scarton, Carolina

A2 - Batista, Fernando

A2 - Silva, Diego

A2 - Magro, Catarina

A2 - Pinto, Hugo

PB - Springer Science and Business Media Deutschland GmbH

Y2 - 21 March 2022 through 23 March 2022

ER -

José MM, José MA, Mauá DD, Cozman FG. Integrating Question Answering and Text-to-SQL in Portuguese. In Pinheiro V, Gamallo P, Amaro R, Scarton C, Batista F, Silva D, Magro C, Pinto H, editors, Computational Processing of the Portuguese Language - 15th International Conference, PROPOR 2022, Proceedings. Springer Science and Business Media Deutschland GmbH. 2022. p. 278-287. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-98305-5_26

Integrating Question Answering and Text-to-SQL in Portuguese

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this