Log in


Scientific creativity

Computational creativity can be used for modeling new scientific discoveries in various fields, such as medicine, biology, mathematics, etc. 

One of the approaches is based on literature-based discovery, where new domain (or better said cross-domain) hypotheses are formed based on identifying interesting "bisociative" terms or links at the intersection of two domains. In closed discovery, experts define two domains and search for new knowledge at their intersection. The seminal work is Swanson's  model (1990), in which literature-based discovery lead to the hypothesis that the lack of magnesium causes migraines. The discovery of bridging terms (so called "b-terms" was explored in e.g. Juršič et al. 2012). On the other hand in "open discovery" the domains are not defined in advance. The semi-automatic definition of second domain was investigated in the context of PubMed articles in the work of Petrič et al. (2009).

A completely different aspect of scientific discovery is explored in our recently developed RoboCHAIR prototype (Pollak et al. 2015), in which we investigate the possibility of a creative assistant, i.e. a system, which takes as an input a scientific article and by questioning the decisions and approaches used by authors, mimic the work of conference chairs and help authors in considering creative alternative approaches in their work. 

The first prototype system is available here: http://kt-robochair.ijs.si/ and will be presented at the IEEE SSCI in December.

In our system users can upload the articles they write, review, read, chair and evaluate the relevance of automatically generated questions. At this stage, the system is working on sentence level, but we aim at developing it further to really produce "creative" questions, by considering knowledge in the article and related work.

If you evaluate the system, but also  any comments that you can upload will be of great help. Moreover, there is a place where reviewers' questions for specific article can be uploaded which would enable us to learn positive examples of appropriate questions. 

For mathematical creativity other researchers (e.g. Alison Pease, Joe Corneli, ...) could explain how to use the potential of computational creativity.


Jursic, M., Cestnik B., Urbancic, T., Lavrac, N. (2012). Finding Bridging Concepts with CrossBee. ICCC 2012.

Pollak,S., Lesjak, B., Kranjc, J.,Podpečan, V., Žnidaršič, M. and Nada Lavrač (2015). RoboCHAIR: Creative Assistant for Question Generation and Ranking. Proc. of IEEE SSCI (to appear).

Petric, I., Urbancic, T., Cestnik B. and Macedoni-Luksic (2009) Literature mining method RaJoLink for uncovering relations between biomedical concepts. J. Biomed. Inform. 42(2): 219-227.

Swanson, D.R. 1988. Migraine and magnesium: Eleven neglected connections. Perspectives in Biology and Medicine 31(4): 526–557

Swanson, D.R. 1990. Medical literature as a potential source of new knowledge. Bull. Med. Libr. Assoc. 78(1): 29–37.

Swanson, D.R.; Smalheiser, N.R.; and Torvik, V.I. 2006. Ranking indirect connections in literature-based discovery:

The role of Medical Subject Headings (MeSH). J. Am. Soc. Inf. Sci. Tec. 57(11): 1427-1439.

Add new comment

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.