Workplan
WP1: User modeling and personalization
This WP aims at building the user modeling and personalization module, through (a) well-designed questionnaires regarding the possible alteration of the user’s musical preferences during COVID-19 and (b) the extraction of features from users’ listening history, which will be correlated with the results of (a).
Task 1.1 User Recruitment
Task 1.2 Sentiment shift analysis
Task 1.3 Feature Extraction and Analysis from User Playlists
WP2: Core Music Information Retrieval Technologies
The objectives of WP2 are to: i) Explore feature representations for all the core modules – music emotion recognition, music similarity and music source separation – included in the music recommendation system. ii) Build machine and deep learning models for the various modules.
Task2.1 Literature review and data specification
Task 2.2 Music emotion recognition
Task 2.3 Music source separation
Task 2.4 Music signal processing for music similarity
WP3: Smart personalized music recommendation system
The objectives of this WP3 are to: i) develop the final smart personalized music recommendation and playlist generation system, ii) validate the developed algorithms through perceptual user tests and iii) employ results of the sentiment shift analysis for refinement and retraining of the developed models.
Task 3.1 Perceptual user tests
Task 3.2 Music Recommendation
WP4: Project Management, Dissemination and Exploitation
Task 4.1 Management and Administration
Task 4.2 Dissemination & Exploitation