Below you can read the papers of the past Conferences
New Music Concepts 2018
Author(s): Anthony Chmiel , Emery Schubert
Abstract: This paper tests the validity of a novel proposal intended for automated music recommendation systems, based on repeated exposure information to the same stimulus contained in a music library. The proposed function is based on a linear combination of Ebbinghaus' forgetting curves which, we argue, produced Berlyne's inverted-U model. It has the effect of a stronger recommendation for a specific music library item after a moderate amount of exposure to that item, and therefore may aid in avoiding over-exposure. The function was tested using two sets of simulations: one set using fixed time increments between exposures, and the other using changing time increments between exposures. All simulations produced inverted-U trajectories, thereby providing evidence that the simple memory principle proposed by Ebbinghaus can be applied to music preference in a parsimonious way.
Keywords: aesthetics, automated music recommendation, exposure, music preference