A Wayuunaiki Translator to Promote Ethno-Tourism Using Deep Learning Approaches.

The Guajira region in Colombia is renowned by ethno-tourists, who immerse themselves in a unique cultural experience by interacting with the Wayúu indigenous peoples. Such tourism is deemed to bring benefits to the region. To further foster tourism, we implement a Spanish-Wayuunaiki translator to facilitate communication between visitors and the Wayúu. We used the well-established CRISP-DM methodology for data mining projects. As a first step, we collect a Spanish-Wayuunaiki corpus of data from a variety of sources. This corpus is used to train the deep learning model employed, which is a recurrent neural network which implements an autoencoder with an attention-based mechanism. Our intention is to integrate the symbolic elements of the Wayúu culture with the traditional activities of ethno-tourism and offer it as a commercial product that provides a unique experience to the visitor. In this way, the natural and cultural resources available in the region are leveraged. Together with the participation of the Wayúu community, it is sought that tourism be an instrument to preserve culture. We implement an application has been deployed in Angular as a frontend on the GitHub Pages service. This application consumes the service deployed in the Heroku Docker Instance, resulting in the application with a translation service on the web. This work aims to contribute to the promotion of the Wayúu culture. Our software is available under an open-source model and approach may be replicated for other indigenous languages.