Department of Mathematics and Computer Science

Context

Leveraging the crowd for idea generation has attracted a lot of attention recently, especially with the emergence of ideation platforms such as Quirky (www.quirky.com) and OpenIDEO (www.openideo.com). These platforms introduced new challenges, for example, many ideas generated by the crowd are mundane and repetitive. It is economically unfeasible to filter out low quality ideas manually due to the high volume of ideas. Therefore, improving the creativity of ideas provided for these platforms has become a focus of research in recent years.

Keywords: Collaborative Ideation, Creativity Support Tools, Intelligent User Interfaces

Problem

One way to improve creativity in open innovation challenges is to show inspiring ideas of others (so called exemplars) to the users while they are generating ideas. This introduces two new questions:

1. When to show exemplars to the user?
2. Which exemplars should be shown to the user?

Related Work deals with this subject (see (1) and (2)) but doesn't include an active model of the user in the system. Therefore there is a need of a user model in collaborative ideation tasks.

Objectives

In this work, a user model should be designed that takes into account on the two questions mentioned above and defines metrics, triggers and user states that could be used by an intelligent system to provide the right exemplars at the right time. Part of the work should be to validate the assumptions made by related work and find appropriate modeling techniques for the user/system collaboration.

Procedure

• Literature review on user modeling in brainstorming tasks (for example (3), (4) or (5))
• Design an intervention model based on the findings of 1.
• Implement the model and test it in a crowd-worker study, in order to
  a) Validate the results of (2)
  b) Test if including concepts from psychology (3) into the model helps the creative outcome (number of ideas, diversity of ideas)
• Validate the assumed states in the model by building a questionnaire that validates the tested user model:

- Number of times stuck, helpfulness of inspirations, ...?

References

(1) Semantically Far Inspirations Considered Harmful? Accounting for Cognitive States in Collaborative Ideation

(2) Providing timely examples improves the quantity and quality of generated ideas

(3) Nijstad, Bernard A., Wolfgang Stroebe, and Hein FM Lodewijkx. "Cognitive stimulation and interference in groups: Exposure effects in an idea generation task." Journal of experimental social psychology 38.6 (2002): 535-544.

(4) Hoffmann, O. (2016). On Modeling Human-Computer Co-Creativity. In Knowledge, Information and Creativity Support Systems (pp. 37-48). Springer, Cham.

(5) Yuan, S. T., & Chen, Y. C. (2008). Semantic ideation learning for agent-based e-brainstorming. IEEE Transactions on Knowledge and Data Engineering, 20(2), 261-275.