Feb 242015

Evolving Sci Comp Creativity 022415
Innovation has become accepted as central to competitiveness in today’s world, both in new product development and in enhancement of internal processes. Companies struggle with innovation, and there have been numerous attempts to regularize and program it. But the development of truly breakthrough ideas is difficult, and recognizing them when they do arrive can be harder still. We have processes available for vetting ideas and passing them through a series of increasingly selective gateways until they reach the point of usefulness or are discarded altogether. But we do not have good processes for stitching together new ideas and reaching that eureka moment that says a critical new idea has been found.

Some of the ways that ideas are sourced include crowdsourcing, internal suggestions, brainstorming, and the like. There are idea factories employing innovative individuals who apply diverse experience to create an “out of the box” concept. And, there are programs such as TRIZ, an innovation program developed in Russia in 1946 that seek to apply a systemic process to ideation itself, based around principles extracted from patent literature subjected to contradiction, synthesis, and new arrangement. But creation of ideas is forever thwarted by the fact that we don’t really understand the creative process and may, in fact, be generalizing a multitude of processes in a way that makes them impossible to replicate.

The most recent forays into the area of programmatic innovation come from the evolving science of computational creativity, a term created to serve as a central term for a general examination of systemic creativity. This is a “grand collaboration” between the arts and the sciences that has resulted in attempts to apply abstracted principles of creativity to a diverse spectrum of areas ranging from music and poetry to general problem solving. It is supported by leading IT vendors, as well as by organizations such as the Association for Computational Creativity and by academia.

To date, the results of computational creativity have been highly mixed, partly because it underscores the problems of defining creativity across the range of fields in which it is important. Some general principles have been abstracted, but they have proven difficult to apply across all fields and, indeed, what is considered “creative” might well differ according to the territory. Is creativity in jazz the same thing as creativity in scientific research? Looking at the creative process itself is, however, beginning to yield new understanding, and the more recent application of big data to this territory is starting to provide some interesting results.

The most visible attempt to bring big data to bear on computational creativity is through IBM’s Watson program, applying its Jeopardy-winning computational appliance to the development of new food recipes (see “IBM Watson and Bon Appétit Team on New App That Transforms How We Cook“). As with Jeopardy, the result appears somewhat trivial on the surface, but they are extending what we know about creative processes and how they can be applied to a diverse range of problems.

Innovation has long suffered from the same issues as creativity, with a few added ones, such as a distinction between market perception of innovation as driven by targeted communications and actual innovation in true breakthrough product design. Creativity feeds innovation. We can program the processes by which innovations are vetted and selected; can we then program the creativity that leads to new design?

Ultimately, it is likely that computational creativity will lead to new tools for design, and new ways of enhancing human input. It is unlikely to take over the innovation process anytime soon. But, in a world starved for new ideas and creative new uses of energy, technology, and materiel, this is an area that deserves to be given closer attention.

Photo by Jared Tarbell via Creative Commons license.


Brian Dooley

Brian J. Dooley is a Senior Consultant with Cutter Consortium's Data Analytics & Digital Technologies practice. He is an author, analyst, and journalist with more than 30 years' experience in analyzing and writing about IT trends.


  2 Responses to “The Evolving Science of Computational Creativity”

  1. Having managed technical groups all my career (in pharma and in chemistry generally) I have constantly asked myself (and anyone within earshot) “How do you manage creativity?”

    My direct answer has become “One cannot.” What is possible is to create an environment where creativity is more likely to occur than otherwise. Several points always seem to be present in such environments; respect for the ideas of others and a general lack of ego (which is a much more accurate phrase than “teamwork”), acceptance of uncertainty and lack of structure, and enjoyment of the task at hand.

    More practically, and having looked at TRIZ and SixSigma, etc. in the past, the simplest way to frame the issue of managing creativity is to ask “How does my present action or thought – whatever it may be – map to a particular desired outcome?” This question is not humanly answerable in any meaningful way; if one knew the answer, the creative process would be unnecessary.

  2. In my opinion, computational creativity will be the key changer in the next few years. We already have witnessed the advent of “intelligent” machines, which are now able to solve several demanding tasks in our stead. The natural prosecution of this is turning machines into creative generators.

    The task is extremely complicated, but the advance in artificial intelligence techniques as well as our growing understanding of the human creative process are really pushing this field forward. In my research area, for example, there are already a number of systems which can generate some interesting music without the need to recur to the human intervention (e.g., Iamus, EMI).

    It is likely that in the next 10 – 15 years we’ll see an increase in the number of hybrid systems where the creative capacity of humans will be enhanced by the support of a machine. But far in the future, it is not unthinkable to predict the advent of fully functioning computational systems that will provide us with creative outputs.

 Leave a Reply

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>