view Report/chapter1/introduction.tex @ 26:e4bcfe00abf4

Final version of code

\setcounter{page}{1}
\pagenumbering{arabic}
\chapter{Introduction}
Music has accompanied social activities on our daily lives and has influenced the shape of the technology landscape that we have today. Portable media players, mobile device applications or music streaming services enable us the access to a large volume of digital recorded music. This vast range of music tracks might include songs that are relevant or not to a listener, being necessary to develop facilities to bring out appropriate musical pieces to an user.

Recommender systems can be described as engines that guide the users to suitable objects from a large number of options in a particular domain such as books, films or music. The available information of users and items' attributes is analysed and exploited by the recommender systems to produce a list of previously unseen items that each user might find enjoyable. Depending on the analysed data, the design of a recommender can be focused on historical ratings given by users or similarities between the attributes of items that an user already rated.

\section{Motivation}
Due to the available information of relationship between users and items would be sparse, e.g., most part of the users tend to do not give enough ratings, the accuracy of predictions would decrease. Another disadvantage of traditional recommender systems, referred as \textit{cold-start problem}, arises when a new item cannot be recommended until it gets enough ratings, or, equivalently, when a new user does not have any ratings \citep{melville2010recommender}. In order to alleviate the rating sparsity and cold-start problems, there is the motivation to combine two or more recommendation designs into hybrid approaches. 

Deep learning is an approach to artificial intelligence for describing raw data as a nested hierarchy of concepts, with each abstract concept defined in terms of simpler representations. For example, deep learning can describe high-level features of an image of a car such as position, color or brightness of the object, in terms of contours, which are also represented in terms of edges. \citep{Bengio-et-al-2015-Book}  

Inspired in natural evolution of species, Estimation of Distribution Algorithms (EDAs) \citep{larranaga2002estimation} are robust techniques developed during the last decade for optimisation in Statistics and Machine Learning fields. EDAs can capture the explicit structure of a population with a probability distribution estimated from the best individuals of that population.

\section{Aim}
We aim to design and implement a hybrid music recommender to suggest new music tracks that an user would find them appealing and enjoyable. The architecture of our hybrid recommender combines two recommendation techniques. 

%The first technique is collaborative filtering to predict music preferences on the basis of users' information from an online social network (OSN) such as Last.fm,

the second technique is \textit{content-based filtering} where recommendations are produced by computing similarities between representations of content of items that an user


in which  are correlated to compute similarities between them.

Users' information is obtained from the Taste Profile dataset, which is a complementary subset of the Million Song Dataset\footnote{http://labrosa.ee.columbia.edu/millionsong/}. The music library that contains sample audio clips of the rated songs in the Taste Profile dataset is consolidated by fetching audio files using 7digital API.

A convolutional neural network (CNN), which is a deep learning model, is employed to describe each audio file of the music library with a n-dimensional vector, whose dimensions represent music genres.

An Estimation of Distribution Algorithm (EDA) technique is implemented to model user profiles in terms of music genres in order to compare each profile with the vector representation of the audio clips to compute similarities between them. Recommendation is achieved by choosing the clips with highest similarity values.

The evaluation of our hybrid music recommender will be assessed by comparing the prediction accuracy with a traditional content-based recommender p.

%that as automatically as possible analyses the multi-track input audio signals, finds and measures the masking phenomenon and uses Equaliser to reduce or solve the problem.
\section{Outline of the thesis}

The rest of the report is organised as follows: Chapter 2 provides an overview in recommender systems. Recommendation process, associated challenges, and related work based on state-of-the-art techniques are discussed. In Chapter 3, we present our proposed hybrid recommendation approach and describe the stages and algorithms in detail. The experiments and evaluation protocols are to assess the performance of the hybrid recommender presented in Chapter 4. We proceed to discuss and analyse the results from the conducted experiments to evaluate the proposed hybrid music recommender. In Chapter 6, we present the conclusions and some thoughts for further research.