Creative exploration using topic-based bisociative networks

Abstract

Bisociative knowledge discovery is an approach that combines elements from two or more "incompatible" domains to generate creative solutions and insight. Inspired by Koestler's notion of bisociation, in this paper we propose a computational framework for the discovery of new connections between domains to promote creative discovery and inspiration in design. Specifically, we propose using topic models on a large collection of unstructured text ideas from multiple domains to discover creative sources of inspiration. We use these topics to generate a Bisociative Information Network-a graph that captures conceptual similarity between ideas-that helps designers find creative links within that network. Using a dataset of thousands of ideas from OpenIDEO, an online collaborative community, our results show usefulness of representing conceptual bridges through collections of words (topics) in finding crossdomain inspiration. We show that the discovered links between domains, whether presented on their own or via ideas they inspired, are perceived to be more novel and can also be used as creative stimuli for new idea generation.

Natural language processing in-and-for design research

“…Zhang et al (2017, p. 2) group 930 concepts (described as paragraphs) that were obtained from a human-centred design course 20 using Word2Vec and the hierarchical clustering algorithm. Ahmed and Fuge (2018, p. 11,12/30) measure topic level association for 3918 ideas that were submitted to OpenIDEO 21 using a Topic Bison Measure, which indicates if a topic pair co-occurs in an idea as well as the proportions of the pair.…”

Communication patterns in engineering enterprise social networks: an exploratory analysis using short text topic modelling

“…Specifically, topic modelling has been used in product design for studying the impact of various interventions on design (Fu, Cagan & Kotovsky 2010; Gyory, Kotovsky & Cagan 2021 a ), to comparing human and AI teams (Gyory et al 2021 b ), to analysing capstone team performance (Ball, Bessette & Lewis 2020), to visualising engineering student identity (Park et al 2020), to deriving new product design features from online reviews (Song et al 2020; Zhou et al 2020) and identifying areas for cross-domain inspiration (Ahmed & Fuge 2018). By plotting a design team’s topic mixtures before and after a manager intervention, Gyory et al (2021 a ) were able to measure the result of the design intervention and whether it helped to bring the team back on track.…”

A Bisociated Research Paper Recommendation Model using BiSOLinkers

“…Moreover, to solve that problem, Amami et al [32] proposed a model where research papers and users were subjected to language and topic modeling respectively to determine the relationship between users and papers and based on the determined closeness of the language used in research papers, unseen research papers were retrieved. Ahmad and Fuge [33] used topic modeling to ensure that the right set of words and topics were utilized to bridge two unrelated domains, achieving contiguity of creative solutions through mediation, similarity, and serendipity [34]- [36]. Lastly, Ahmad and Fuge [33] used topic models to discover topical links that bridge two unrelated domains.…”

“…Ahmad and Fuge [33] used topic modeling to ensure that the right set of words and topics were utilized to bridge two unrelated domains, achieving contiguity of creative solutions through mediation, similarity, and serendipity [34]- [36]. Lastly, Ahmad and Fuge [33] used topic models to discover topical links that bridge two unrelated domains. Their proposed model was a computational framework for discovering new connections while supporting creativity and the discovery of novel and new ideas.…”