Semantic Interaction With Music Content Using Foaf

Semantic Interaction with Music Content using
FOAF
`
Oscar
Celma1 , Miquel Ramrez1 , and Perfecto Herrera1
Music Technology Group, Institut Universitari de lAudiovisual
Universitat Pompeu Fabra
Ocata 1, 08003 Barcelona, Spain.
{ocelma, mramirez, pherrera}@iua.upf.es
http://www.iua.upf.es/mtg
Abstract. SIMAC (Semantic Interaction with Music Audio Contents)

is an European Commission project (IST-2.3.1.7, Semantic-based knowledge systems) which aims to develop a set of prototypes to describe,
semi-automatically, music audio content. SIMAC is about music metadata, about what you can say of a piece of music, on what is hidden in a
music file, in a collection of music files, and in the collective knowledge of
a community of music lovers. This document proposes to enhance/extend
the FOAF definition to model user musical tastes. One of the goals of
the project is to explore music content discovery, based on both user
profiling and content-based descriptions.
Introduction
The main goal of SIMAC1 is doing research on semantic descriptors of music

contents, in order to use them, by means of a set of prototypes, for providing song
collection exploration, retrieval and recommendation services. These services are
meant for home users, music content producers and distributors and academic
users. One special feature is that these descriptions are composed by semantic
descriptors. Music will be tagged using a language close to the users own way
of describing its contents moving the focus from low-level to higher-level (i.e.
semantic) descriptions.
SIMAC is about music metadata. We assume that metadata is all what one
can say about a single music piece or a collection of music pieces. The collective
knowledge of a community of people interested in music is meant also to be
handled as metadata.
The Semantic Web and Music Metadata
The World Wide Web has become the host and distribution channel of a broad
variety of digital multimedia documents. Although the Internet infrastructure
1
http://www.semanticaudio.org
`
Oscar
Celma et al.
allows simple, straight-forward acquisition, the value of these resources suffers

from a lack of powerful content management, retrieval and visualization tools.
Music content is no exception: although there is a sizeable amount of textbased information about music (album reviews, artist biographies, etc.) this
information is hardly associated to the objects they refer to, that is music pieces.
In the context of the Semantic Web, there is a clear interest to create a
Web of machine-readable homepages describing people, the links between them
and the things they create and do. The FOAF (Friend Of A Friend ) project2
provides conventions and a language to tell a machine the sort of things a
user says about herself in her homepage. FOAF is based on the RDF/XML3
vocabulary. As we noted before, the knowledge hold by a community of peers
about music is also a source of valuable metadata and FOAF nicely allows users
to speak about whatever they fancy. Music is an important vehicle for telling
other people something relevant about our personality, history, etc.
There already exist some approaches4 to modeling music in the FOAF context. A simple way is to depict interests in an artists works as shown in example
2.1.
<foaf:interest
rdf:resource="http://www.norahjones.com"
dc:title="Norah Jones" />
Example 2.1: Using FOAF to express interest of an artist
In this example, Dublin Core definitions are used to express the artists title. A simple question arise: What is Norah Jones?. Is it a book title? Obviously, a country and bluesy music fan will not mistake the Norah Jones term
as a title of a book, but a software agent will. The dc:title tag definition is being
overloaded. In other words, a machine would not understand that the user is
interested in a well-known singer and composer.
Already built-in within FOAF, there are ways to say that a user is interested
in a topic (see example 2.2). Even though there is no taxonomy of topics, this
example gives more general information than the one shown in example 2.1.
There is an interesting project, SongBuddy5 , that offers a way to find music
available on the Internet by browsing peoples FOAF profiles. By finding songs
on bands and labels sites and sharing the URLs of those songs with other users,
music discovery becomes possible. Apart from the usual FOAF information, it
is required to specify a set of songs as RDF resources. SongBuddy describes this
kind of resources using the MusicBrainz6 RDF definition. Associated metadata
2
3
4
5
6
http://www.foaf-project.org
http://www.w3.org/RDF
All examples shown here have been gathered from the web
http://www.songbuddy.com
http://www.musicbrainz.org
Semantic Interaction with Music Content using FOAF
<foaf:topic_interest>
<rdf:Description>
<dc:subject>Music</dc:subject>
<dc:description>Pop</dc:description>
</rdf:Description>
</foaf:topic_interest>
Example 2.2: Using FOAF to express interest in a topic

<foaf:Person rdf:nodeID="nickname">
<foaf:mbox_sha1sum>
ce24ca1400c2f511c6s4b015a1f064dda8356f9a
</foaf:mbox_sha1sum>
<mm:trackList>
<rdf:Seq>
<rdf:li
rdf:resource="http://www.agivensite.org/holeinmysoul.mp3"/>
<rdf:li
rdf:resource="http://www.anothersite.com/music/Thirteen.mp3"/>
</rdf:Seq>
</mm:trackList>
</foaf:Person>
<mm:Track
rdf:about="http://www.agivensite.org/holeinmysoul.mp3">
<dc:title>Hole in my Soul</dc:title>
<dc:creator
rdf:resource="http://www.songbuddy.com/lc/soaf/artist/name=Aerosmith"/>
</mm:Track>
<mm:Track rdf:about="http://www.anothersite.com/music/Thirteen.mp3">
<dc:title>Thirteen</dc:title>
<dc:creator
rdf:resource="http://www.songbuddy.com/lc/soaf/artist/name=Amy+McKenna"/>
</mm:Track>
<mm:Artist rdf:about="http://www.songbuddy.com/lc/soaf/artist/name=Aerosmith">
<dc:title>Aerosmith</dc:title>
</mm:Artist>
<mm:Artist rdf:about="http://www.songbuddy.com/lc/soaf/artist/name=Amy+McKenna">
<dc:title>Amy McKenna</dc:title>
</mm:Artist>
(...)
Example 2.3: Using FOAF in SongBuddy project
for a song are track and artist name. An example of a Songbuddys FOAF
definition is shown in example 2.3. Unfortunately, SongBuddy does not offer any
automated system that can propose new musical assets a user could be interested
in.
`
Oscar
Celma et al.
Our proposal
As we said in the introduction, SIMACs primary concern is the development

of effective music content discovery tools. We are already accumulating a significant wealth of knowledge about music not only trivial editorial data, but
knowledge such as relationships between artists, musical works and also song
descriptions automatically extracted by means of signal processing and machine
learning techniques. On the other hand, we are conscious that practical content
discovery systems must also have some knowledge about users. We see FOAF
as an excellent candidate for becoming a framework to model peoples musical
taste.
Currently, we are devising our own music ontology in OWL7 . We can foresee
that embedding/integrating these musical concepts within FOAF will allow our
system to better understand user questions. Example 3.1 shows an hypothetical
FOAF profile using our approach. This approach allows to express what a user
likes or dislikes regarding to music items, with fine-grained detail from a genre,
artist, album to a track level. A user can also annotate reviews about a song
or an album. The system will process all this information when discovering new
music content. Figure 1 shows an overview of the system.
It is likely that a selection of relevant items that are purely based on content
descriptors such as pitch, tempo, etc. will either return too many hits or
silly answers. Hence, transforming the data about the user (these data become
information once they have been processed) defines what makes sense and what
does not, from that users point of view.
3.1
Expected results
Merging (or embedding) a music ontology within FOAF will allow us to understand a user in two complementary ways; psychological factors extracted
from the normal FOAF profile personality, demographic, socio-economics,
situation and explicit musical preferences. Using this enhanced FOAF profile, the system will be able to filter and to contextualize users queries.
This approach opens a wide range of possible usages and applications, such
as:
Recommendation of new musical pieces, supported by both content-based
and context-based similarity criteria
Recommendation of unknown artists playing close to a users location
whose works are similar to users music taste
Finding, automatically, new mates that have similar musical tastes, building
an adaptive, evolving and ever-growing social network
http://www.w3.org/TR/owl-features
Semantic Interaction with Music Content using FOAF
<foaf:Person rdf:nodeID="nickname">
<foaf:mbox_sha1sum>
ce24ca1400c2f511c652b015a1k76f9sa13dk6f9a
</foaf:mbox_sha1sum>
<simo:dislikes>
<simo:genre simo:name="Classical Music" />
<simo:soundslike>
<simo:object>
<simo:artist simo:name="Aerosmith" />
</simo:object>
</simo:soundslike>
<simo:soundslike>
<simo:object>
<simo:track simo:title="Heart of glass">
<simo:artist simo:name="Blondie" />
</simo:track>
</simo:object>
</simo:soundslike>
</simo:dislikes>
<simo:likes>
<simo:artist simo:name="P.J. Harvey" />
<simo:album simo:title="More unchartered heights of disgrace">
<simo:artist simo:name="The Dogs DAmour" />
</simo:album>
</simo:likes>
<simo:review>
<simo:Patient simo:type="track" simo:name="Missed" simo:rating="9" />
<simo:Text lang="en">
I like very much this PJ Harveys song because guitars sound wild (...)
</simo:Text>
</simo:review>
</foaf:Person>
Example 3.1: Simac project approach. Hypothetical FOAF profile
`
Oscar
Celma et al.
Fig. 1. System overview using FOAF profile.

Semantic Interaction With Music Content Using Foaf

Uploaded by

Copyright:

Available Formats

Semantic Interaction With Music Content Using Foaf

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Semantic Interaction With Music Content Using Foaf

Uploaded by

Copyright:

Available Formats

Semantic Interaction with Music Content using

Abstract. SIMAC (Semantic Interaction with Music Audio Contents)

The main goal of SIMAC1 is doing research on semantic descriptors of music

The Semantic Web and Music Metadata

allows simple, straight-forward acquisition, the value of these resources suffers

Example 2.1: Using FOAF to express interest of an artist

Semantic Interaction with Music Content using FOAF

Example 2.2: Using FOAF to express interest in a topic

Example 2.3: Using FOAF in SongBuddy project

As we said in the introduction, SIMACs primary concern is the development

Semantic Interaction with Music Content using FOAF

Example 3.1: Simac project approach. Hypothetical FOAF profile

Fig. 1. System overview using FOAF profile.

You might also like