Information Technology Reference
In-Depth Information
the visual representation of real-world objects and scenes (in addition, of course,
to imaginary ones). We hope to build models which can subvert this information in
playful and productive ways, building meaningful scenes different to those it was
given.
We also intend to extend the collage generation approach described above,
whereby online resources are employed as art materials. To this end, we have begun
construction of a Computational Creativity collective, available at:
www.doc.ic.ac/
ccg/collective
.
The collective currently contains individual processes which perform
creative, analytical, and information retrieval tasks, along with mashups, which
combine the processes in order to generate artefacts of cultural interest. For in-
stance, the project whereby news stories are turned into collages described above is
modelled in the collective as a mashup of five processes which retrieve news sto-
ries, extract text, retrieve images and construct collages. The collective currently has
processes which can link to Google, Flickr, the BBC, LastFM, Twitter and numer-
ous other online sources of information. Our plans for the collective are ambitious:
we hope to attract researchers from various areas of computing including graph-
ics, natural language processing, computer music and audio, and AI to upload their
research systems to expand the collective.
Systems built for Computational Creativity purposes such as The Painting Fool
are beginning to have abilities of note in their particular domains of expertise, but
rarely are they combined in order to increase the cultural value of their output. Hence
we plan to paint pictures using the text produced by story generators like the Mexica
system (Perez y Perez
2007
) as input, and there is no reason why the pictures pro-
duced couldn't be used, for instance, as input to an audio generation system. This
example highlights the masterplan for the collective, which is to have the output
of one system continually consumed as the input to another system, thus provid-
ing a framework for experimentation with control mechanisms. In particular, the
first mechanism we will implement will be based on Global Workspace Architec-
tures, (Baars
1988
), as per the PhD work of Charnley (
2010
). It is our hope that the
increase in complexity of processing, coupled with the ability to access culturally
important information from online resources will lead to more thought-provoking
artefacts being generated.
An important part of our future research will be to continue to engage audi-
ences on an artistic level, i.e., by organising exhibitions and reacting to the feed-
back we gain from such exercises. As an example, in April 2011, we exhibited art-
works from The Painting Fool alongside those by traditional artist Eileen Chen, who
worked in watercolours and graphic pens. The exhibition was entitled “No Photos
Harmed/Growing Paths from Seed”, and was a dialogue in which we explored the
handing over of creative responsibility in artistic processes. In traditional painting
approaches, with the subject matter and more pointedly with mediums such as wa-
tercolours, the artist has to occasionally go with the flow, hence doesn't retain full
creative responsibility. We tried to emphasise the continuation of this with Compu-
tational Creativity projects, whereby such responsibilities are explicitly and wilfully
handed over to software.
