p@36: 
p@36: \section{Introduction}
p@36: 
p@36: Recommender systems are software facilities for providing items suggestions or predicting customer behavior by using prior user information. These systems play an important role in commercial applications to increase sales and improve user satisfaction.\\
p@36: \\
p@36: We introduce a hybrid music recommender to mitigate the \emph{cold-start} problem. Our approach is inspired on the results in ~\cite{DBLP:journals/corr/KereliukSL15} and ~\cite{Liang2014781} using convolutional neural networks (CNN) for music genre classification and estimation of distribution algorithms (EDA) for user modeling. Our findings support the idea that a combination of techniques might
p@36: improve the recommendation performance.
p@36: %\begin{itemize}
p@36: %\item Our aim is to perform automatic fruit classification.
p@36: %\item $AFC$ is vital for a number of surveillance applications such as %\emph{temporal decomposition} and automatic \emph{fruit collision} detection.
p@36: %\item As the number of fruit increases we must design our system with scaleability.
p@36: %\end{itemize}