Kaleidoscope: Computer-Assisted Ideation through Idea Network Exploration

Idea generation (ideation) is an essential process in human creativity that leads to great innovation. Brainstorming is one of the most widely-used and powerful ideation methods. In this paper, we propose a novel approach to assisting human brainstorming, which allows computers to collaborate with humans and boost new ideas. To achieve this goal, we first model the brainstorming process in a network structure, then construct idea networks using heterogeneous datasets varying in both domains and forms. We define objectives to discover insightful ideas from idea networks, and design algorithms that satisfies these objectives. As an outcome, we develop an online brainstorming system with automated idea recommendation and evaluate our approach by user surveys. To the best of our knowledge, our work is the first attempt to enable human-computer collaboration in an interactive ideation process. Our contribution lies in the following: (1) we initiate the methodology of computerassisted ideation; (2) we craft objectives and algorithms that shed insight on recommending knowledge in information networks; (3) we build a system to facilitate ideation with knowledge and evaluate its effectiveness.