Implementing Vocal Natural Language Interface to Enterprise Resource Planning System - Université de technologie de Compiègne
Proceedings Year : 2023

Implementing Vocal Natural Language Interface to Enterprise Resource Planning System

Shengzhe Zhang
  • Function : Author

Abstract

In order to meet the challenge of industry in economic globalization, enterprises around world have built information systems to effectively manage the production, achieve integration of suppliers and customer and organize all the resources by software. However, the interaction between the system and users are not always intuitive and some of the questions about production cannot easily to be answered in today’s Enterprise Resource Planning (ERP). This paper aims at proposing a vocal natural language interface which adopts text-to-SQL neural networks for an open source ERP system Odoo for purpose of interacting with the complex ERP system easily by voice. The current automatic speech recognition can transform voice into text with reliability. For the next step of text-to-SQL, although there is no large amount of data of natural-language-query and SQL pairs in production, the neural text-to-SQL parser can be trained on less specific databases and data and worked on specific closed domain databases like ERP. Moreover, this cross-domain feature of neural text-to-SQL parser makes it possible to build such an interface. The result showed that such an interface can be applied on actual usage scenario and informs the user of the manufacturing status. The proposed system is now implemented on command line and a graphic interface will make it more intuitive.
No file

Dates and versions

hal-04395089 , version 1 (15-01-2024)

Identifiers

Cite

Shengzhe Zhang, Julien Le Duigou. Implementing Vocal Natural Language Interface to Enterprise Resource Planning System. International Joint Conference on Mechanics, Design Engineering and Advanced Manufacturing, JCM 2022, Springer International Publishing, pp.399-409, 2023, Lecture Notes in Mechanical Engineering, ⟨10.1007/978-3-031-15928-2_35⟩. ⟨hal-04395089⟩
17 View
0 Download

Altmetric

Share

More