ALIA-FM, an automatic and intelligent prototype of a music radio presenter:

Smart assistants

Overview

ALIA-FM is an automatic and intelligent prototype of a music radio presenter. It was developed as a demo to demonstrate the validity of the models created in the ALIA project, with the aim of automatically creating music radio shows in Basque and Spanish. The tool selects the songs, the information about them and their context, and generates a coherent voice-over based on the criteria set by the user via a web interface.

Challenge

The main challenge was to build a system capable of automatically generating a radio show that goes beyond a simple reading of the metadata of the songs. The system had to contextualise the songs and the artists, establish links between them and construct a natural and continuous commentary between songs, in both Basque and Spanish.

Collaboration

It was developed in collaboration with the EHU.

Result

The result of the project is the functional demo ALIA-FM, which allows users to configure the knowledge bases to be used, as well as the theme, language, number of presenters, and gender of the voices, to create a commentary-driven music show. Its architecture combines the engine for selecting the songs, the generation of the commentary and creation of the voice-over using the LLM and TTS model developed in the ALIA project. The prototype code can be found at GitHub: https://github.com/orai-nlp/alia-fm