A Personal Perspective into the Future for Computational Creativity - Computational Creativity Research: Towards Creative Machines - page 394

computationally if one extended the framework so that search could range over the
possible traversal and evaluation functions, as well as the conceptual spaces defined
by each such choice. With respect to the processes described in Sect. 19.2 , such an
extension of the search would correspond to instantiation of the design processes.
This has therefore been known to be a theoretical possibility but has only been con-
sidered in implemented systems in a very limited fashion, as described above (see
Sect. 19.2.3 ). If Wiggins was right, future generations of computationally creative
systems will include more and more of the design processes, resulting in an improved
perception of their creativity.
Colton et al. [ 3 ] postulated Computational Creativity Theory as a term to encom-
pass efforts to describe generic characteristics of what it implies for a computer to
be creative. To my knowledge, the term has not been adopted by a wider community
and remains applicable only to their FACE and IDEA descriptive models. The moti-
vation described for these models is “to suggest future directions for computational
creativity research”. The models presented by Colton et al. had the merit of bringing
attention to a number of issues that have indeed given rise to valuable research efforts.
The conceptual schema described in this chapter could be considered a contribution
along similar lines. The relation with the framing component of the FACEmodel, and
how the conceptual schema presented here goes beyond it to describe the reflection
process in relation to all the others has been described in Sect. 19.2.3 .
Veale [ 20 ] campaigns in favour of a future for Computational Creativity as a
connected network of creative web services. He argues that such a service-oriented
architecture would allow researchers to pool creative technologies in a robust inter-
operable framework, and industry developers to exploit novel results of CC research
in a robust, low-risk form, without having to re-implement algorithms from a quickly
moving field. With respect to the set of processes described above, it is clear that
functions representing the selection criteria subject of processes 3 and 4, and the
constructive procedures subject of process 1, could be very well implemented as
web services. Web services could also be good vehicles for implementing providers
of ingredients subject of process 2. In each case, if a number of alternative versions
where available for each service, software components that implemented decisions
of which particular versions to use would constitute instantiations of the processes
themselves. In such a setting, processes 5 and 6—applying a constructive procedure to
ingredients and selecting valuable results out of the constructed candidates—could
be developed by composition of simpler services so selected. It is less clear how
process 7 of reflection on the results could be implemented in such an architecture,
but it might take the form of a service broker, explicitly handling the interconnections
between the corresponding services. One notable exception is the likelihood of instan-
tiating process 8 for establishing the novelty of particular results. By its nature, such
a process requires the consideration of state—in terms of the set of artefacts of the
same kind already contemplated. This is contrary to the definitions of a service most
often considered, which includes statelessness as one of its requirements.
Varshney et al. [ 19 ] argue along the lines that a creative system without a human
cognitive component for the purpose of evaluation is severely handicapped. This
matches the prediction by Bringsjord and Ferrucci [ 2 ]. If one accepts this prediction,