Building datasets for automated conversational systems designed for use-cases
Juliana Duarte de Camargo, J.M.B. Nunes, Maria João Antunes, Óscar Mealha, Carolina Abrantes, Luis Nobrega · 2022
The efficiency of an Artificial Intelligence (AI) based system depends on the richness and accuracy of its training dataset. These datasets are responsible for training the algorithms used by conversational systems, making them more effective and able to respond autonomously to user interactions. When it comes to specific contexts and communicational situations, there are still gaps in building datasets that are shaped to different realities. Therefore, the present study sought to develop a dataset to support the creation and training of an intelligent conversational system that enables autonomous care, via telephone calls, of restaurant, beauty salon and medical clinic's customers. The stages of subjective user experience data collection, preparation and validation are reported.