XRDS • SPRING 2018 • VOL. 24 • NO. 3
sulting style representation can then
be applied to photographic images,
like holiday snapshots or portraits of
friends and family, to create new images in the style of a particular artist
or genre of painting.
Finally, new kinds of user experiences and human-computer interfaces are creating spaces for co-creation
between humans and intelligent algorithms that push the boundaries
of the human mind. When machine
intelligence is integrated into human
workflows, the sources of creativity are
mingled and the demarcation line bet ween human and machine is blurred.
As you can tell, I dream of a future
where people and machines work hand
in hand to enable our species to make
quantum leap innovations and solve
the pressing problems our society is
facing. We already see the benefits of
such technologies in health, resource
management, business applications,
and in the arts—for many the last bastion of human creativity. But gladly
some artists, like Mario Klingemann,
don’t see themselves in competition
with technology but let it work for
them, even inspire them (see Figure
1). To him technologizing creativity
doesn’t take anything away from its
humanity. On the contrary, in his opinion an iterative interaction between
human and computer can be highly
stimulating for novel, creative thought.
As Mario demonstrated at a recent Sonophilia meeting devoted to creative
AI, his algorithms help him explore
combinations of possibilities, which in
turn trigger new ideas and give way to
new, intriguing artworks [ 11].
Now that we’ve broken the myth and
brought creativity back to the ordinary
mortal, we can start technologizing it.
Creativity is not a sudden flash but
a brick-by-brick process, and as such it
can be modeled, learned, and taught.
And this makes creativity—at least theoretically—emulatable by machines.
They can help us enhance our level of
expertise and knowledge by enabling
us to digest huge chunks of information that would otherwise be impossible to gather in a lifetime. The convergence of various technologies and
methods that have matured in recent
years play a central role here.
Today we have the ability to build
large data repositories that function as
equivalents to memory or experience
banks in our human neuronal networks. These data collections allow us
to systematically explore data and the
information that can be derived from
it. Most importantly, however, they
form the basis of the learning systems
we can use to teach computers how to
understand the world around them.
In recent years, systems for building
and training artificial neural networks
have matured and applications of these
networks on huge data sets have resulted in sometimes spectacular success.
These so-called deep-learning systems
have been successfully applied to beat
humans at chess or Go, but also to identify objects in photos, moving images
and even sound. Basically, these systems extract characteristics from large
collections of data and detect patterns.
It is very difficult to point to the exact
features that are detected; this idea was
written about by Jürgen Schmidhuber
in 1992. Schmidhuber, a leading computer scientist in artificial intelligence,
describes these neural networks as
landing on a description of ideas and
concepts on a subconscious level [ 9].
Based on these learnings, style
transfer algorithms and special configurations of neural net works can apply and contextualize abstract ideas
and concepts to concrete manifestations of new creations. One prime
example of this is “learned representation for artistic style” that shows
particular stylistic traits of artists can
be learned from data derived from
the artworks themselves [ 10]. The re-
What if creativity
was not reserved for
the select few but
What if it were not
ACM’s Interactions magazine
explores critical relationships
between people and
emerging innovations and
industry leaders from around
the world across important
applications of design thinking
and the broadening ;eld of
Our readers represent a growing
community of practice that is
of increasing and vital global
To learn more about us,
visit our award-winning website
Follow us on
Facebook and Twitter