A hybrid approach for artwork recommendation

IGNACIO GATTI

Sardinia

Conferencia; 2019 IEEE Second International Conference on Artificial Intelligence and Knowledge Engineering (AIKE); 2019

Università degli Studi di Cagliari

In this article, we give an overview of the particular-ities that surround artwork Recommender Systems in museums.Unlike other domains, such as movies, where the user selectsonly one movie to watch, in the museum domain it is necessaryto recommend a sequence of items to consume, in which the orderand relationships among them might affect the recommendation.Furthermore, curators determine a place for each artwork inthe museum, generating a distribution over the museum?s roomsthat imposes constraints about the way visitors walk through it.However, nowadays it is also possible to access to their digitalcollection, gaining independence from the physical distribution,allowing any visitor to be his/her own curator. Because of that, inthis article we analyze and discuss different approaches to exploitthe particularities of this domain. Specifically, we focus on oneaspects, namely: item modeling. First, we summarize differentapproaches to represent and extract features from artworksthat could be relevant for the recommendation task. Finally, wepropose a novel approach to model artworks that could be usefulin the recommendation task.