Informatica 42 (2018) 69–76 69
Computational Creativity in Slovenia
Senja Pollak1, Geraint A. Wiggins2,3, Martin Žnidaršič and Nada Lavrač1,4
1 Department of Knowledge Technologies, Jožef Stefan Institute, Ljubljana, Slovenia
2 Computational Creativity Lab, Queen Mary University of London, London E1 4NS, UK
3 AI Lab, Free University of Brussels, Brussels 1050, Belgium
4 University of Nova Gorica, Nova Gorica, Slovenia
E-mail: {senja.pollak,martin.znidarsic,nada.lavrac}@ijs.si; geraint.wiggins@qmul.ac.uk
Keywords: computational creativity, bisociative reasoning, computational creativity platform
Received: November 6, 2017
Computational Creativity is a field of Artificial Intelligence that addresses processes that would be deemed
creative if performed by a human. The field has been very active since 1999, and is now an established
research field with its own International Conference on Computational Creativity (ICCC) conference series
founded in 2010. This paper briefly surveys the field of Computational Creativity (CC) that is based on
the analysis of ICCC conference papers, followed by a more detailed presentation of projects and selected
contributions of Slovenian researchers to the field.
Povzetek: Računalniška ustvarjalnost je področje umetne inteligence, ki obravnava procese, ki bi jih oce-
nili kot kreativne, če bi jih izvajal človek. Področje računalniške ustvarjalnosti se je razmahnilo po letu
1999, kot veja znanosti pa se je uveljavilo leta 2010 z ustanovitvijo serije letnih konferenc z imenom In-
ternational Conference on Computational Creativity (ICCC). V članku podamo kratek pregled področja
računalniške ustvarjalnosti, ki temelji na analizi ICCC konferenčnih člankov, posebno pozornost pa name-
nimo predstavitvi projektov in izbranih dosežkov slovenskih raziskovalcev.
1 Introduction
As a sub-field of Artificial Intelligence (AI) research, Com-
putational Creativity1 (CC) is concerned with machines
that exhibit behaviours that might reasonably be deemed
creative [49; 11]. Slovenian researchers have made impor-
tant contributions to CC. This paper aims to provide an ob-
jective snapshot of the field of computational creativity as
a whole, and to give a brief summary of the particular con-
tribution of Slovenian researchers to it.
In the next section, we summarise an analysis of the re-
search field, that we conducted in 2016, using it to structure
a brief introduction to the field for unfamiliar readers. We
then summarise the contributions of Slovenian researchers
to CC.
2 Domain understanding
We here summarise the results of a study of the research
field of Computational Creativity [36], which was based
on the analysis of papers published in the Proceedings of
the International Conference on Computational Creativity
(ICCC)2. The aim of the study was to objectively identify
areas of interest in this research field. Here, we use its con-
clusions to motivate our subsequent outline of CC research.
1http://computationalcreativity.net
2http://computationalcreativity.net/home/
conferences/
In the previous study, Pollak et al. [36] used semi-
automatic topic ontology generation tool OntoGen [16] to
explore the texts of the complete conference proceedings of
the International Conference on Computational Creativity
to date. This allowed them to make an objective, explaina-
ble bottom-up analysis of the field.
The input to the OntoGen tool are documents, which
are texts of individual articles from the proceedings. Af-
ter manual text cleaning and removal of the papers’ refe-
rence sections, OntoGen performs stemming and stop word
removal, followed by the construction of Bag-of-Words
(BoW) feature vector representations of documents, where
the features are weighted by the TF·IDF heuristic [41] and
used for clustering. The user may explore the results, and
identify hierarchies of significant terms and clusters of do-
cuments. The keywords are identified by OntoGen in two
ways: descriptive keywords are extracted from document
centroid vectors, while distinctive keywords are extracted
from the SVM classification model distinguishing the do-
cuments in the given topic (document cluster) from the do-
cuments neighbouring clusters [16]. Other functionalities
used were expert’s manual moving of documents between
clusters to reduce inappropriate paper categorisation and
active learning of selected concepts/categories.
Several outputs were presented by Pollak et al. [36], in-
cluding understanding of the field of Computational Crea-
tivity based on its topics, which is also of interest to this
study.
A final corpus-based categorisation of the field of com-
70 Informatica 42 (2018) 69–76 S. Pollak et al.
Figure 1: Semi-automatically generated conceptualization of the CC domain, with CC concept naming and sub-concept
creation.
putational creativity is presented in Figure 1. The main
sub-domains of computational creativity identified by our
method were: Musical, Visual, Linguistic creativity, Ga-
mes and creativity, Conceptual creativity as well as ne-
wly created category of Evaluation. For several domains,
subcategories were detected also at lower levels, including
Narratives, Poetry, Recipes and Lexical creativity as subdo-
mains of Linguistic creativity. Each category can be further
characterised through descriptive keywords listed in Table
1, as extracted from cluster centroid vectors.
3 Brief review of computational
creativity
We now present a brief overview of Computational Crea-
tivity, as represented by the domains identified by Pollak
et al. [36]. We have added in an additional category, scien-
tific creativity, on the grounds that important work in this
area was performed prior to the inception of the ICCC con-
ference, and was therefore not represented in the conferen-
ces.
In this position paper we do not present a detailed re-
view of the field but explain the key issues and cite some
successful exemplars of CC research. A recurring general
theme of ICCC is the attempt to better understand what is
meant by term “creativity.” Early on, it was recognised that
we must move away from Romantic notions of “great” cre-
ativity, if we are to make progress. So ICCC is interested
in creative process more than creative output, and there is
no acceptance of the notion of “inspiration”, understood as
mystical intervention by some agency extrinsic to the cre-
ator. Of course, in a paper such as this, one cannot discuss
process without reference to outputs, without being inter-
minably dull. For this reason, we include examples where
possible.
Boden [1, 2] first formally raised the question of creati-
vity in AI, but there have been significant precursors of CC
field in several domains that are also mentioned here.
3.1 Visual creativity
Most work on visual creativity is conceptualised in terms
of painting or drawing. In this domain, there tends to be a
focus on painting technique and on the objects produced.
The clear forerunner of CC in this domain was Harold
Cohen, a successful artist in his own right, who built a ro-
bot painter, AARON3, programmed in a rule-based style.
Its development began in the 1970s, with developments
right up to the artist’s death in 2016 [26; 2]. Cohen vie-
wed AARON as a part of his art, and therefore did not al-
ways disclose the methods used to make it work, though
he did write several papers on some aspects of the system
[e.g., 7; 6; 5]. Figure 2(a) shows a well-known painting by
AARON.
Simon Colton’s The Painting Fool4 deconstructs painting
from subject composition (for example, collage based on
stories from The Guardian newspaper) right down to brush
stroke [9]. Figure 2(b) shows an example.
DARCI5 [27], unusually, is multi-modal and can explain
itself: it combines image processing with language com-
prehension, so as to focus the system on the extraction and
generation of meaning. DARCI produced the image in Fi-
3www.aaronshome.come
4http://www.thepaintingfool.com
5http://darci.cs.byu.edu
Computational Creativity in Slovenia Informatica 42 (2018) 69–76 71
Table 1: Categories and keywords of the first layer of the semi-automatically constructed CC ontology.
Category Automatically extracted keywords
Musical music, chord, improvisation, melodies, harmonize, composition, accompaniment, pitch, emotions, beat
Visual image, painting, darci, artifacts, collage, adjectives, associations, rendered, colored, artists
Linguistic story, poems, actions, character, words, agents, narrative, artefacts, poetry, evaluating
Games games, design, player, games design, angelina, agents, code, jam, filter, gameplay
Conceptual analogy, blending, mapping, conceptual, objective, associations, team, graphs, concepts, domain
Evaluation music, poems, improvisation, evaluating, interactive, poetry, creativity system, musician, participants, beha-
vioural
Comp. creativity music, image, story, games, agents, words, actions, poems, character, blending
gure 2(c), explaining it as follows (there is not space here
for the intermediate images): “I looked at this picture, [an
elephant walking across a verdant African plain] and it re-
minded me of this image that I’ve seen before, [a standing
stone] which is a picture of a stone. The picture also see-
med gloomy and brooding. So I created this initial sketch,
[black and white graphic drawing] and then rendered it in a
style related to stone, gloomy, and brooding, which resul-
ted in this image. [intermediate image] It turned out more
like a bucket or a cauldron, and it seems creepy, but I’m
happy with it.”
3.2 Creative game design
Computational creativity has many applications in games,
perhaps most obviously in the area of game level genera-
tion, where the landscape and structure of a game are cre-
ated live. However, probably the most unexpected and in-
teresting example of CC in games is Yavalath6 [3], ranked
in the top 100 board games ever invented by the Board-
GameGeek website. It is highly novel in that the board is
hexagonal.
Another success has been Angelina7 [12], a long term
project aiming for completely computational creativity of
digital games.
3.3 Linguistic creativity
Creativity in language covers a broad area, including poetry
and story-telling. Two systems that demonstrate different
approaches are MEXICA [31] and Propper [17]. MEXICA
uses a general creative method, the Engagement-Reflection
model, to model a two-phase, cyclic approach to creativity.
Propper takes a contrasting approach, using heuristics from
literary theory [38] to guide exploratory creativity. A third
successful approach is that of Tony Veale [46]. Veale’s lab
specialises in the development of elegant methods of ex-
tracting data from lingustic corpora, and then using that
data for creative text generation, often in TwitterBots—see
@MetaphorMagnet [e.g., 47].
6http://www.cameronius.com/games/yavalath/
7http://www.gamesbyangelina.org
3.4 Musical creativity
Musical creativity had important precursors too. David
Cope’s EMI [13; 50] produced many compositions, but
none of the reports on the work made it clear what the
system actually did, and how much was due to its author.
A clearer early contribution, with full scientific reporting,
was by [14], which produced musical harmony in the style
of J. S. Bach. This is a remarkable contribution, and still
stands today as an excellent piece of work; its fault is that
its harmonisations sound too much like Bach—the system
does not reflect on its overall balance, but applies Bachian
compositional tricks everywhere.
Perhaps the first attempt at automated composition really
to situate itself in CC was the work of [28]. Melodies were
generated from a learned model of style, and evaluated in
detail by expert musicians [29].
François Pachet’s team has produced the most thorough
CC music systems to date, working from chords and melo-
dies right through to studio production [40].
3.5 Scientific creativity
It is often forgotten that human creativity is evident in
science and engineering, as well as in the arts and huma-
nities. One of the earliest successes in CC was the HR
system of Colton [8]. This was an exploratory creativity
system, which invented new integer sequences with pro-
perties that mathematicians find interesting; 17 of the se-
quences it dicovered were novel and interesting enough to
be included in the Journal of Integer Sequences, which re-
cords these structures and acts as an encyclopedia of them.
It also made conjectures about some of these sequences that
were subsequently proven correct.
Another successful project in scientific creativity was
funded by the EU FP7 programme: BISON studied the ap-
plication of bisociative reasoning [20] to medical text ana-
lysis (see Section 4.1).
3.6 Concept creation
Concept creation arises as a separate category in the ob-
jective analysis because it is central to all creative domains.
72 Informatica 42 (2018) 69–76 S. Pollak et al.
(a) (b) (c)
Figure 2: Three computationally created images. (a) Untitled from AARON’s middle period output. (b) The Painting
Fool’s Uneasy. (c) DARCI’s Always Be A Gloomy Cauldron, Even in Creepy Stone.
There are too many approaches to survey here; however,
a recurring theme is conceptual blending [44], which has
been carried forward with some success. An example is the
Divago system [30], a computational model that uses con-
ceptual blending. The key idea here is somewhat similar to
Koestler’s bisociation [20]: new concepts are created from
combinations of features of existing and/or imagined ones.
A recent EU FP7 project, ConCreTe, focused on Concept
Creation Technology (see Section 4.3).
3.7 Creative systems evaluation
Evaluation is a particularly difficult problem in comptua-
tional creativity, which attracts commensurate attention in
the literature. There are two distinct ways that computatio-
nally creative systems involve evaluation: first, in the con-
ventional scientific sense, where the correctness and value
of work is assessed; and, second, in the sense of reflection
within the system, that allows it to make intelligent creative
decisions. Quite often, but certainly not always, these two
aims coincide.
The value of a creative act is a function of four aspects
[51]: Context, Observer, Creator and Artefact, forming the
acronym COCA. But this does not give detail of how cre-
ativity might actually be assessed. Ritchie [39] gives a de-
tailed set of criteria that can be used to assess the creativity
of a computer program, which have been used in several
projects. Jordanous [e.g., 18] and van der Velde et al. [e.g,
45] have made substantial contributions in this area.
4 Computational creativity in
Slovenia
To the best of our knowledge, the only Computational Cre-
ativity research in Slovenia has been performed by the
members of Department of Knowledge Technologies at
Jožef Stefan Institute (JSI) in Ljubljana. Most of the rese-
arch, including the work summarised in Section 2, has ta-
ken place within three distinct EU-funded projects and the
PROSECCO networking action, all supported by the Euro-
pean FP7 funding programme. We summarise this work,
with a special focus on Slovenian contributions.
4.1 Bisociation networks for creative
information discovery (BISON)
BISON8 was a research project from the field of scientific
creativity, which deals with the bisociation-based scienti-
fic knowledge discovery. Arthur Koestler [20] argued that
the essence of creativity lies in “perceiving of a situation
or idea . . . in two self-consistent but habitually incompa-
tible frames of reference”, and introduced the expression
bisociation to characterise this creative act. The key vi-
sion of the BISON project was to develop a fundamentally
new ICT paradigm for bisociative information discovery.
JSI’s main contributions were related to scientific literature
mining aimed at creatively forming new hypotheses based
on yet uncovered relations between knowledge from diffe-
rent, relatively isolated fields of specialization. We deve-
loped CrossBee9, a literature-based discovery support tool
[19], where different elementary and ensemble heuristics
8http://cordis.europa.eu/project/rcn/86374_en.
html
9http://crossbee.ijs.si/
Computational Creativity in Slovenia Informatica 42 (2018) 69–76 73
are implemented for bisociative bridging term (b-term) dis-
covery. The heuristics are defined as functions that numeri-
cally evaluate the term quality by assigning it a bisociation
score (measuring the potential that a term is actually a b-
term). Other methodologies developed for cross-domain
literature based discovery focus on exploration of outlier
documents [34; 42]. JSI’s methods were tested on standard
datasets (e.g., migraine-magnesium studied in early rese-
arch by [43], but also actually led to new hypotheses in
understanding autism [23] and Alzheimer’s disease [4].
4.2 The What-If Machine (WHIM)
The WHIM project was concerned with the automated ge-
neration, understanding and evaluation of fictional ideas.
Fictional ideas are propositions of situations that are un-
realistic or commonly considered as unplausible, such as:
“What if there was a little fish who couldn’t swim?” which
are a central part of various creative works and products.
Artificial production of What-if ideas is creative work that
is inherently hard to automate, but there are now some ge-
nerators available (e.g., [22]). In the generation process,
there is usually a trade-off between a template driven pro-
cess (with a relatively narrow covering of the fictional ide-
ation space) or more open and autonomous generative pro-
cess (producing more interesting and valuable ideas, but
larger amount of lower quality results).
The What-if Machine was also the inspiration for a real
musical show Beyond The Fence, billed as “the world’s
first computer-generated musical”, that performed in Lon-
don in 2016. In this artistic project—containing the mu-
sical and a documentary—several computational creativity
research prototypes were combined and used in the artistic
process [10].
JSI’s main role in the WHIM project was in automated
modelling of human evaluations. The main tasks included
the design of a large crowd-sourcing data gathering exer-
cise, resulting in more than 10,000 evaluated fictional ideas
and next, to build data mining models, which would al-
low differentiation between the sentences, appreciated by
human evaluators as good/creative (regarding their novelty
and narrative potential) or bad. We tested also an alter-
native approach for gathering human evaluations through
interaction with the robot Nao [35]. Other contributions of
Slovenian researchers to the WHIM project included biso-
ciative generation of fictional ideation [32] and the Robo-
Chair10 system for enhancing scientific creativity by gene-
rating questions regarding decisions made by authors when
writing scientific articles [37].
4.3 Concept creation technology
(ConCreTe)
The ConCreTe11 project focused on AI technology for con-
cept construction, identification, and evaluation. ConCreTe
10http://kt-robochair.ijs.si/
11http://www.conceptcreationtechnology.eu
addressed several forms of conceptual blending (CB), a ba-
sic cognitive mechanism by which two or more mental spa-
ces are integrated to produce new concepts [15]. Optima-
lity principles (OPs), a key element in the CB framework,
are responsible for guiding the integration process towards
good blends. The role of OPs was studied from the point
of view of computational systems [24], as well as within a
study of human perception of visual animal blends12 [25],
performed with the aim of better understanding of creative
artefacts reception.
The main contribution of JSI to ConCreTe was the Con-
CreTeFlows platform13 [48] for collective CC workflows
construction. It is a platform built on top of the exis-
ting ClowdFlows infrastructure [21], but it is specialised at
supporting (primarily text-based) computational creativity
tasks, such as conceptual blending and poetry generation. It
currently contains more than 35 native widgets for suppor-
ting creativity by developers from five different institutions
participating in ConCreTe. The asset of a web-based sy-
stem is that it integrates creative software written in a large
variety of programming languages (e.g., components writ-
ten in Python, C#, Java, PROLOG). An interesting example
of multimodal conceptual blending [48] is available as an
interactive workflow.14
4.4 Other projects and activities
We have described the main projects from the field of CC
in which we were actively involved. Other project were
closely related to computational creativity. For example,
within the EU project MUSE15, the question of interactive
story-telling was addressed. Our main role was the integra-
tion of the developed components in the online workflow
environment [33].
The PROSECCO16 networking action had a crucial role
in building the European CC community, with a number
of events including the organisation of summer schools,
code camps, etc. Computational Creativity has became
an important research topic in Slovenia. A large number
of activities were organised also by Slovenian researchers
and held place in Ljubljana, including the 5th edition of the
ICCC conference17, with material available through Vide-
oLectures18, and the Symposium on Computational Crea-
tivity19. We have also organised the Computational Creati-
vity art exhibition entitled You/Me/It.20
Since 2016, a Computational Creativity course has been
offered at the International Postgraduate School Jožef Ste-
12http://animals.janez.me/
13http://concreteflows.ijs.si
14http://concreteflows.ijs.si/workflow/137/
15http://www.muse-project.eu/
16http://prosecco-network.eu/
17http://procsecco-network.eu
18http://videolectures.net/iccc2014_ljubljana/
19http://videolectures.net/ktsymposium2013_
ljubljana/
20http://computationalcreativity.net/iccc2014/
you-me-it-art-exhibition/
74 Informatica 42 (2018) 69–76 S. Pollak et al.
fan21.
As CC related outreach activity, a large number of events
for children and youth were organised for science promo-
tion by means of a Nao robot, for which the main developer
Vid Podpečan received the Slovenian “Prometej znanosti”
(Prometheus of Science) science dissemination award.
5 Conclusion
This paper presented a brief review of historic and current
activity in Computational Creativity, an exciting and rela-
tively new sub-field of Artificial Intelligence. In particular,
we have highlighted contributions from Slovenian resear-
chers.
Computational Creativity is in some sense a final frontier
for AI [11], because it pulls the field away from comforta-
bly defined problem-solving activity such as classification,
into the areas that are more challenging to formulate. Much
of the work in this developing field is focused not so much
on “What is the answer?” but rather on “What is the ques-
tion?”, and this makes for exciting prospects for the future,
both in Slovenia and elsewhere. In 2008, the Association
for Computational Creativity22 (ACC) was founded to ma-
nage the ICCC conferences and support the CC community
into the future.
Acknowledgements
We acknowledge the support of the Slovenian Research
Agency (core funding no. P2-0103), the European pro-
jects Prosecco (grant no. 600653) and ConCreTe (grant
nb. 611733). GW is very grateful to the International Pos-
tgratuate School Jožef Stefan internationalisation grant for
funding a sabbatical visit in Autumn 2017, which enabled
his contribution to this paper.
Literature
[1] Boden, M. (1977). Artificial Intelligence and Natural
Man. Harvester Press.
[2] Boden, M. A. (2004). The Creative Mind: Myths and
Mechanisms (2nd ed.). Routledge.
[3] Browne, C. (2008). Automatic Generation and Evalu-
ation of Recombination Games. Ph. D. thesis, Queens-
land University of Technology.
[4] Cestnik, B., E. Fabbretti, D. Gubiani, T. Urbančič,
and N. Lavrač (2017). Reducing the search space in
literature-based discovery by exploring outlier docu-
ments: A case study in finding links between gut mi-
crobiome and alzheimers disease. Genomics and Com-
putational Biology 3(3), 58.
21https://www.mps.si/splet/studij.asp?lang=eng&
main=1&left=4&id=721&m=4
22http://computationalcreativity.net
[5] Cohen, H. (1979). What is an image? In Proceedings
of the 1979 International Joint Conference on Artificial
Intelligence.
[6] Cohen, H. (1988). How to draw three people in a bo-
tanical garden. In Proceedings of the 1988 Conference
of the American Association for Artificial Intelligence
(AAAI-88).
[7] Cohen, H. (1999). Colouring without seeing: A pro-
blem in machine creativity. AISB Quarterly 102, 26–35.
[8] Colton, S. (2012a). Automated Theory Formation in
Pure Mathematics. Distinguished Dissertations. Sprin-
ger London.
[9] Colton, S. (2012b). The painting fool: Stories
from building an automated artist. In J. McCormack
and M. d’Inverno (Eds.), Computers and Creativity.
Springer-Verlag.
[10] Colton, S., M. T. Llano, R. Hepworth, J. W. Charn-
ley, C. V. Gale, A. Baron, F. Pachet, P. Roy, P. Gervás,
N. Collins, B. L. Sturm, T. Weyde, D. Wolff, and J. R.
Lloyd (2016). The Beyond the Fence musical and Com-
puter Says Show documentary. In Proceedings of the Se-
venth International Conference on Computational Cre-
ativity, UPMC, Paris, France, June 27 - July 1, 2016.,
pp. 311–321.
[11] Colton, S. and G. A. Wiggins (2012). Computational
creativity: The final frontier? In de Raedt L. et al. (Ed.),
Proceedings of ECAI Frontiers.
[12] Cook, M., S. Colton, A. Raad, and J. Gow (2013).
Mechanic miner: Reflection-driven game mechanic dis-
covery and level design. In A. I. Esparcia-Alcázar (Ed.),
Applications of Evolutionary Computation: 16th Euro-
pean Conference, Proceedings, pp. 284–293. Springer.
[13] Cope, D. (1992). Computer modelling of musical in-
telligence in EMI. Computer Music Journal 16(2), 69–
83.
[14] Ebcioğlu, K. (1988). An expert system for harmoni-
zing four-part chorales. Computer Music Journal 12(3),
43–51.
[15] Fauconnier, G. and M. Turner (2002). The Way We
Think. New York: Basic Books.
[16] Fortuna, B., D. Mladenič, and M. Grobelnik (2006).
Semi-automatic construction of topic ontologies. In Se-
mantics, Web and Mining: Joint International Works-
hops, EWMF 2005 and KDO 2005, Revised Selected
Papers, pp. 121–131. Springer.
[17] Gervás, P. (2015). Computational drafting of plot
structures for Russian folk tales. Cognitive Computa-
tion.
Computational Creativity in Slovenia Informatica 42 (2018) 69–76 75
[18] Jordanous, A. (2012). A standardised procedure for
evaluating creative systems: Computational creativity
evaluation based on what it is to be creative. Cognitive
Computation 4(3), 246–279.
[19] Juršič, M., B. Cestnik, T. Urbančič, and N. Lavrač
(2012, may). Cross-domain literature mining: Finding
bridging concepts with crossbee. In Proceedings of the
Third International Conference on Computational Cre-
ativity, Dublin, Ireland, pp. 33–40.
[20] Koestler, A. (1976). The Act of Creation. London,
UK: Hutchinson.
[21] Kranjc, J., V. Podpečan, and N. Lavrač (2012). Clo-
wdFlows: A cloud based scientific workflow platform.
In Machine Learning and Knowledge Discovery in Da-
tabases - European Conference, ECML PKDD 2012,
Bristol, UK, September 24-28, 2012. Proceedings, Part
II, pp. 816–819.
[22] Llano, M. T., S. Colton, R. Hepworth, and J. Gow
(2016). Automated fictional ideation via knowledge
base manipulation. Cognitive Computation 8(2), 153–
174.
[23] Macedoni-Lukšič, M., T. Urbančič, I. Petrič, and
B. Cestnik (2016). Autism research dynamic through
ontology-based text mining. Advances in Autism 2(3),
131–139.
[24] Martins, P., S. Pollak, T. Urbančič, and A. Cardoso
(2016). Optimality principles in computational appro-
aches to conceptual blending: Do we need them (at)
all? In Proceedings of the Seventh International Confe-
rence on Computational Creativity (ICCC 2016), Paris,
France. Sony CSL: Sony CSL.
[25] Martins, P., T. Urbančič, S. Pollak, N. Lavrač, and
A. Cardoso (2015). The good, the bad, and the aha!
blends. In Proceedings of ICCC, pp. 166–173. compu-
tationalcreativity.net.
[26] McCorduck, P. (1991). AARON’S CODE: Meta-Art,
Artificial Intelligence and the Work of Harold Cohen’S
CODE: Meta-Art, Artificial Intelligence and the Work of
Harold Cohen. Freeman.
[27] Norton, D., D. Heath, and D. Ventura (2013). Fin-
ding creativity in an artificial artist. Journal of Creative
Behavior 47(2), 106–124.
[28] Pearce, M. T. (2005). The Construction and Evalua-
tion of Statistical Models of Melodic Structure in Music
Perception and Composition. Ph. D. thesis, Department
of Computing, City University, London, London,UK.
[29] Pearce, M. T. and G. Wiggins (2007). Evaluating cog-
nitive models of musical composition. In A. Cardoso
and G. Wiggins (Eds.), Proceedings of the 4th Internati-
onal Joint Workshop on Computational Creativity, Lon-
don, pp. 73–80. Goldsmiths, University of London.
[30] Pereira, F. C. (2007). Creativity and Artificial Intelli-
gence: A Conceptual Blending Approach. Berlin: Mou-
ton de Gruyter.
[31] Pérez y Pérez, R. and M. Sharples (2001). Mexica: A
computer model of a cognitive account of creative wri-
ting. Journal of Experimental & Theoretical Artificial
Intelligence 13(2), 119–139.
[32] Perovšek, M., B. Cestnik, T. Urbančič, S. Colton, and
N. Lavrač (2013). Towards narrative ideation via cross-
context link discovery using banded matrices. In IDA,
Volume 8207 of Lecture Notes in Computer Science, pp.
333–344. Springer.
[33] Perovšek, M., V. Podpečan, J. Kranjc, T. Erjavec,
S. Pollak, N. Q. Do Thi, X. Liu, C. Smith, M. Cavazza,
and N. Lavrač (2015). Text mining platform for NLP
workflow design, replication and reuse. In Proceedings
of IJCAI Workshop on Replicability and Reusability in
Natural Language Processing: From Data to Software
Sharing, Buenos Aires, Argentina, 26 July 2015.
[34] Petrič, I., B. Cestnik, N. Lavrač, and T. Urbančič
(2012, January). Outlier detection in cross-context
link discovery for creative literature mining. Comput.
J. 55(1), 47–61.
[35] Podpečan, V. (2015). The What-If machine robot in-
terface (WHIMBOT). In Show, tell imagine: A day
to explore computational creativity together, pp. 17.
Queen Mary, Univ. of London.
[36] Pollak, S., B. M. Boshkoska, D. Miljkovic, G. Wig-
gins, and N. Lavrač (2016). Computational creativity
conceptualisation grounded on iccc papers. In V. C. F.
a. G. François Pachet, Amilcar Cardoso (Ed.), Procee-
dings of ICCC 2016, pp. 123–130. Association for Com-
putaitonal Creativity.
[37] Pollak, S., B. Lesjak, J. Kranjc, V. Podpečan,
M. Žnidaršič, and N. Lavrač (2015). RoboCHAIR: Cre-
ative assistant for question generation and ranking. In
Proceedings of SSCI, pp. 1468–1475. IEEE.
[38] Propp, V. (1968). Morphology of the folktale. Austin:
University of Texas Press.
[39] Ritchie, G. (2007). Some empirical criteria for at-
tributing creativity to a computer program. Minds and
Machines 17(1), 67–99.
[40] Sakellariou, J., F. Tria, V. Loreto, and F. Pachet
(2017). Maximum entropy models capture melodic sty-
les. Scientific Reports 7(9172).
[41] Salton, G. and C. Buckley (1988). Term-weighting
approaches in automatic text retrieval. Information Pro-
cessing & Management 24(5), 513–523.
76 Informatica 42 (2018) 69–76 S. Pollak et al.
[42] Sluban, B., M. Juršič, B. Cestnik, and N. Lavrač
(2012). Exploring the Power of Outliers for Cross-
Domain Literature Mining, pp. 325–337. Berlin, Hei-
delberg: Springer Berlin Heidelberg.
[43] Swanson, D. R., N. R. Smalheiser, and V. I. Torvik
(2006). Ranking indirect connections in literature-based
discovery: The role of medical subject headings. Jour-
nal of the American Society for Information Science and
Technology 57(11), 1427–1439.
[44] Turner, M. and G. Fauconnier (1995). Conceptual in-
tegration and formal expression. Metaphor and Symbo-
lic Activity 10(3), 183–203.
[45] van der Velde, F., R. Wolf, M. Schmettow, and D. Na-
zareth (2015, 6). A semantic map for evaluating creati-
vity. In H. Toivonen, S. Colton, M. Cook, and D. Ven-
tura (Eds.), Proceedings of the Sixth International Con-
ference on Computational Creativity (ICCC 2015), pp.
94–101. WordPress. Open access.
[46] Veale, T. (2012). Exploding the Creativity Myth. New
York, NY: Bloomsbury Academic.
[47] Veale, T. and G. Li (2016, Apr). Distributed divergent
creativity: Computational creative agents at web scale.
Cognitive Computation 8(2), 175–186.
[48] Žnidaršič, M., A. Cardoso, P. Gervás, P. Martins,
R. Hervás, A. O. Alves, H. G. Oliveira, P. Xiao,
S. Linkola, H. Toivonen, J. Kranjc, and N. Lavrač
(2016). Computational creativity infrastructure for on-
line software composition: A conceptual blending use
case. In Proceedings of the Seventh International Con-
ference on Computational Creativity, UPMC, Paris,
France, June 27 - July 1, 2016., pp. 371–379.
[49] Wiggins, G. (2006). A preliminary framework for
description, analysis and comparison of creative sys-
tems. Journal of Knowledge Based Systems 19(7), 449–
458.
[50] Wiggins, G. (2007). Models of musical similarity.
Musicae Scientiae 11, 315–338.
[51] Wiggins, G., P. Tyack, C. Scharff, and M. Rohr-
meier (2015). The evolutionary roots of creativity: me-
chanisms and motivations. Philosophical Transactions
of the Royal Society of London B: Biological Scien-
ces 370(1664).