In this paper, a new intent and entity recognition model for the subject area of air passenger service, labelled as IRERAIR-TWIN, is developed using the ‘no code’ question-answer development platform ‘TWIN’. The advantages of the no-code platform were analysed in terms of the ease of developing an application question-answer system and reducing the amount of work involved in developing an application model for a narrow subject area. The results show that the ‘TWIN’ system provides an intuitive web-based user interface and a simpler approach to develop the semantic module of a question-answer system capable of solving application problems for a narrow subject area that are not overly complex. However, this approach has limitations for deep semantic analysis tasks, especially in complex contextual inference and processing of large text fragments. The paper concludes by emphasising that future research will focus on using ChatGPT-based ‘low code’ platforms and large language models to further improve the intelligence of the IRERAIR-TWIN model. This extension aims to broaden the scope of the scenarios.

Keywords: question-answering systems, No-code, Low-code, Intent recognition, Named entity recognition, Data annotation, Feature engineering, Pre-trained model, software development,End-user development

This paper explores in detail the technological evolution and current state of question and answer (Q&A) systems. Using an example of an airline customer service task, a BERT-based model is developed that is capable of recognising user intentions and extracting named entities. The paper provides a detailed description of the dataset preparation, data analysis methods and data exploration techniques of the project. A description of the model and parameter settings during the model tuning process and the model training process is presented. The model developed in this project is named RNEEMAviCS-BERT, which achieved an intent recognition accuracy of 98.2% and named entity recognition accuracy of 83%. We have created a semantic analysis module for the question and answer system. The next stage of our work will be to integrate the dataset to complete the query-response and response generation components of the Q&A system.

Keywords: question-answering systems, ChatGPT, BERT, machine learning, neural networks, pre-trained models, intention recognition, named entity recognition, data analysis, model training