What is Computational Creativity?

Share

Synopsis

What is computational creativity? This post explores the seeming oxymoron of combining computation and creativity, via the search for a definitive definition of computational creativity.

In my previous post, could you tell that I rather hastily skipped over something...? It’s all very well talking about computational creativity, its achievements, its development, its research community - but what on earth is computational creativity?

An attempt to answer that question deserves more attention than a quick definition in an introductory article. So now, let’s give this question our full attention.

According to the Computational Creativity Conference Steering Committee (the group behind many computational creativity research events):

“Computational creativity is a multidisciplinary endeavour that is located at the intersection of the fields of artificial intelligence, cognitive psychology, philosophy and the arts.

The goal of computational creativity is to model, simulate or replicate creativity using a computer, to achieve one of several ends:

• to construct a program or computer capable of human-level creativity

• to better understand human creativity and to formulate an algorithmic perspective on creative behavior in humans

• to design programs that can enhance human creativity without necessarily being creative themselves

The field of computational creativity concerns itself with theoretical and practical issues in the study of creativity. Theoretical work on the nature and proper definition of creativity is performed in parallel with practical work on the implementation of systems that exhibit creativity, with one strand of work informing the other.“ (See  more computationalcreativity. net)

This definition of computational creativity has been developed and built up over years of discussion and disputing, rewording and reworking, and probably a fair share of sweating and swearing as well. There are, inevitably, issues with this current definition - some pedantic, some trivial, some troubling (I leave it to you to decide which issues fall into which categories):

  •  What is meant by “human level creativity”? The creativity of a genius? A child’s creativity? Where is this bar set?
  •  Should the computational work achieve more than “one of” these “ends, does that mean it is doubly creative? or not creative at all? (OK that’s definitely one of the most pedantic issues...)
  •  What if the computational work helps us to “better understand human creativity but not from an “algorithmic perspective on creative behaviour” (perhaps it uses evolutionary techniques to “evolve” creative behaviour, for example) - is this therefore not computational creativity?
  •  Similarly, if an “algorithmic perspective on creative behaviour” produces results we might feel inclined to describe as creative, but this perspective doesn’t really shed any light on human creativity, does this necessarily exclude it from being computational creativity?
  •  If computational creativity can be computational work that enhances human creativity, then aren’t we really dealing with computer-enhanced human creativity, not computational creativity per se?
  •  If the purpose of theoretical work in computational creativity is to look at “the nature and proper definition of creativity”, does this mean that we can’t actually define computational creativity (because we are still working on what "creativity" is)?
  •  What is the relationship between computational creativity and human creativity, anyway?

Perhaps the most pressing issue, though, is that this “definition” doesn’t directly answer the question “what is computational creativity?”. Instead, it rephrases the question as three new questions:

  1.  What disciplines does computational creativity cover?
  2.  What are the goals of computational creativity?
  3.  What does computational creativity research entail?

In some ways, such rephrasing is to be praised - the original question is tremendously awkward to answer. But there used to be a definition at computationalcreativity.net which tried to address the original question:

Computational Creativity is the study and simulation, by computational means, of behaviour, natural and artificial, which would, if observed in humans, be deemed creative.” 

This seemed more like a direct description of what computational creativity actually is, but was dropped as the "official" definition during 2013.

A key change to be noted between this definition and the current revision (apart from length and level of detail) is the removal of the suggestion of a computer appearing to be human - this concept has been very carefully and deliberately reworded in the current definition (Colton & Wiggins 2012).

Is this change to make computational creativity more generally acceptable and less philosophically challenging? Perhaps. The whole idea of computational creativity may otherwise seem a little odd, or even paradoxical, depending on how comfortable you feel about the idea of a computer being creative.

On the other hand, is it possible to conceive of a creativity that is in some way distinct from human creativity? (NB Some may argue that such a conception is our very answer to "what is computational creativity", though this viewpoint is a little extreme and is definitely not an interpretation of the current official definition of computational creativity, quoted earlier in this post.)

I believe that the recent changes to the definition are to make the implications of the word "creativity" more general, rather than having to be linked to human creativity. Achieving human-level creativity is a lofty aim; relaxing that aim somewhat (even temporarily) gives computational creativity more scope for progress. Otherwise, it’s like asking a newborn Mozart to progress straight onto composing symphonies, without any time to develop creative prowess. (Mozart was 5 years old when he started composing - what took him so long? ... )

Defining computational creativity, "officially" or not, is a task that has been weighing heavily on the minds of people within the computational creativity community for some time. "[A] formal definition of creativity [in the context of computational work] – and our inability to find one that satisfies everybody - has been the elephant in the room at all of the computational creativity workshops to date." (Cardoso et al., 2009)

There have been many contributions from an array of different opinions and perspectives. Let’s look at some of these computational creativity definitional offerings over recent years, listed loosely in reverse chronological order, to get an idea of the evolution of the current definition, in amongst the differences of opinion:

  • "Computational Creativity research is: The philosophy, science and engineering of computational systems which, by taking on particular responsibilities, exhibit behaviours that unbiased observers would deem to be creative." Colton and Wiggins (2012, my emphasis added)
  • Many people defined computational creativity as creativity involving computers in some way. (Is this cheating - to side step the definition of computational creativity by using people's implicit conception of creativity?)
    • "computer programs are written to perform tasks which, if undertaken by people, require an element of creativity." Colton (2008)
    • "Computational creativity is the study and simulation, by computer means, of behaviour, natural and artificial, which would, if observed in humans, be deemed creative." Cardoso and Wiggins (2007)
    • "The study and support, through computational means and methods, of behaviour exhibited by natural and artificial systems, which would be deemed creative if exhibited by humans." Wiggins (2006a)
    • "as long as creativity and its prerequisites are not well understood, it probably is the best approach [for computational creativity] to simulate human behaviour as far as possible." Schmid (1996)
  • [Computational creativity is] "how to create something new and useful at the same time." Peinado and Gervas (2006). Here computational creativity borrows from psychological research into creativity, e.g. Mayer (1999); Plucker et al. (2004). In practice, novelty (related concepts: originality, newness) and value (related concepts: usefulness, appropriateness, relevance, significance) of creative products had often been identified as the two main aspects of creativity to concentrate on in computational work, for a while, almost in tacit agreement. It is questionable, of course, whether creativity can or should - be reduced to just these two components. Life just isn’t that simple for the creativity researcher!
  • “It seems natural to interpret the creative process, particularly in a computational context, as one of search. ” Ventura (2011). The idea of the creative process as “the search for an ideal solution to a problem” (Jennings 2010) has been floating around for quite a few years now, certainly since 2006 (Wiggins 2006b).
  • Pease et al. (2001) suggest redefining creativity as a concept called creativity2 so that it is more tightly defined and therefore easier to assess. Creativity2 is however left undefined by Pease et al....
  • Engagement-reflection (Pérez y Pérez, 1999) is where the system takes a cyclic process of engaging with creative production then critically reflecting on what it has produced, to inform the next stage of creative production. “Engagement-reflection [is] a model used to implement systems to simulate computational creativity” Pérez y Pérez et al. (2010).
  • Cohen (1999) defines a behavior X as the combination of emergence, awareness, motivation and knowledge, and considers his generative art system AARON in the context of behavior X. Cohen refrains, however, from describing AARON as creative - largely because AARON lacks the ability to make aesthetic choices about the quality of produced work. The implication (left implicit by Cohen) is that behavior X and autonomy are two necessary conditions for creativity, though it is unclear whether Cohen sees this combination as sufficient for creativity.
  • Considering computational creativity from a philosophical standpoint, Boden (1990) identifies three different types of creativity: exploratory creativity (exploring a scope-limited set of possible options within a domain - Boden uses the term conceptual space for this set); transformational creativity (transforming the conceptual space (set of options) by identifying where the boundaries of the set can be changed to include new options); Combinational creativity (combining two or more concepts within the conceptual space to form a new concept).

Definitions, definitions everywhere. But, have we found an answer to the question "What is computational creativity?" I doubt it, though we’ve covered a lot of ground in what computational creativity might be, with some (perhaps misguided) pedantism while examining current definitional offerings. As you see here, the answer to our question is not trivial - it is still work in progress.

I have a sneaking suspicion that John McCarthy, the artificial intelligence pioneer, got it right all along when thinking about how various human qualities could be incorporated into a machine. I’ll let him have the final word:

"To ascribe certain beliefs, knowledge, free will, intentions, consciousness, abilities or wants to a machine or computer program is legitimate when such an ascription expresses the same information about the machine that it expresses about a person. It is useful when the ascription helps us understand the structure of the machine, its past or future behavior, or how to repair or improve it." McCarthy (1979)

References

(Many of these references can be found via computationalcreativity.net/resources/books/ and computationalcreativity.net/resources/bibliography/

Boden, M. A. (1990). The creative mind: Myths and mechanisms. Basic Books, Inc, New York.

Cardoso, A., Veale, T., and Wiggins, G. A. (2009). Converging on the divergent: The history (and future) of the international joint workshops in computational creativity. AI Magazine, 30(3):15-22.

Cardoso, A. and Wiggins, G. A., editors (2007). Proceedings of the 4th International Joint Workshop on Computational Creativity, London, UK. IJWCC, Goldsmiths, University of London.

Cohen, H. (1999). Colouring without seeing: A problem in machine creativity. In AISB Quarterly - Special issue on AISB99: Creativity in the arts and sciences, volume 102, pages 26-35.

Colton, S. (2008). Creativity versus the perception of creativity in computational systems. In Proceedings of AAAI Symposium on Creative Systems, pages 14-20.

Colton, S. and Wiggins, G. A. (2012). Computational creativity: The final frontier? In Proceedings of 20th European Conference on Artificial Intelligence (ECAI), pages 21-26, Montpellier, France.

Jennings, K. E. (2010). Search strategies and the creative process. In Proceedings of the International Conference on Computational Creativity, pages 130-139, Lisbon, Portugal.

Mayer, R. E. (1999). Fifty years of creativity research. In Sternberg, R. J., editor, Handbook of Creativity, chapter 22, pages 449-460. Cambridge University Press, Cambridge, UK.

McCarthy, J. (1979). Ascribing mental qualities to machines. In Ringle, M., editor, Philosophical Perspectives in Artificial Intelligence. Humanities Press, Atlantic Highlands, NJ.

Pease, A., Guhe, M., and Smaill, A. (2010). Some aspects of analogical reasoning in mathematical creativity. In Proceedings of the International Conference on Computational Creativity, pages 60-64, Lisbon, Portugal.

Pease, A., Winterstein, D., and Colton, S. (2001). Evaluating machine creativity. In Proceedings of Workshop Program of ICCBR-Creative Systems: Approaches to Creativity in AI and Cognitive Science, pages 129-137.

Peinado, F. and Gervas, P. (2006). Evaluation of automatic generation of basic stories. New Generation Computing, 24(3):289-302.

Pérez y Pérez, R. (1999). MEXICA: A Computer Model of Creativity in Writing. PhD thesis, University of Sussex, Brighton, UK.

Pérez y Pérez, R., Aguilar, A., and Negrete, S. (2010). The ERI-Designer: A computer model for the arrangement of furniture. Minds and Machines, 20(4):533-564.

Plucker, J. A., Beghetto, R. A., and Dow, G. T. (2004). Why isn’t creativity more important to educational psychologists? Potentials, pitfalls, and future directions in creativity research. Educational Psychologist, 39(2):83-96.

Schmid, K. (1996). Making AI systems more creative: The IPC-model. Knowledge-Based Systems, 9(6):385-397.

Ventura, D. (2011). No free lunch in the search for creativity. In Proceedings of the 2nd International Conference on Computational Creativity, pages 108-110, Mexico City, Mexico.

Wiggins, G. A. (2006a). A preliminary framework for description, analysis and comparison of creative systems. Knowledge-Based Systems, 19(7):449-458.

Wiggins, G. A. (2006b). Searching for computational creativity. New Generation Computing, 24(3):209-222.

Article Featured Image: "RoboMunch" by Tony Veale. See also Tony's new illustrated book on computational creativity, available at RobotComix.com

Tags: algorythmic perspective on creative behavior, anna jordanous, computational creativity, creativity measures, creativity research, creativity studies, definition of creativity, maggie boden

blog comments powered by Disqus