Information Technology Reference
In-Depth Information
however, the point of the exercise is to write software to make people think more.
This helps in the argument against people who are worried about automation en-
croaching on intellectual life: in fact, in our version of an AI-enhanced future, our
software might force us to think more rather than less. Note further that there are
also powerful works of art which emphasise the phenomenological experience of
the work, or which are best appreciated through types of meditation. Hence, as well
as hoping to increase mental activity with some of the artefacts that our software
produces—which would literally change peoples' minds, whether in terms of an
long-held opinion or temporary feeling—we should also hope to change the state of
the minds of audience members. In either case, it is clear that, if we want Compu-
tational Creativity software to have impact on people, it should have individual and
collective models of the minds of audience members.
1.4 Illustrative Projects
Our purpose with The Painting Fool project is to build an automated painter which
is one day taken seriously as a creative artist in its own right. In order to do this,
we have developed a roadmap based on the notion of climbing a meta-mountain as
described above. That is, we have specified a sequence of very broad areas within
which to research and implement improved versions of the software, by asking the
question “what does a painter do?” and answering as follows:
1. Makes marks on a canvas
2. Represents objects and scenes pictorially
3. Paints scenes in different styles
4. Chooses styles in a meaningful way
5. Paints new scenes from imagination
6. Invents scenes for a purpose
7. Learns and progresses as an artist.
Naturally, this is a very subjective and quite naive breakdown of painterly pro-
gression, and is not intended for anything other than directing components within
our research programme. As such, it serves its purpose well, and as we will see each
component described below fits into one of the parts of this roadmap and contributes
to the overall goal of producing an independent artist. For each component, our over-
riding aim is to implement more sophisticated versions of the software. However,
determining what represents an improved program is often one of the more diffi-
cult aspects of the project, and we use both engineering standards and feedback
from people who view the artworks produced to assess the level of success of each
component. Hence, for the majority of the components, we present details of the
motivations and aims; some implementation details; results from scientific testing
of the software; and a gallery of images arising from running the software, along
with some commentary on the value of the images, based on the feedback we have
received.
