Blender is a system that combines artificial and human intelligence to increase creativity.
The current version is preliminary and merely produces stimulus phrases associated with a particular concept. Blender has several methods for generating these phrases, all of which are still under development. The methods are in the folder generators.
We hypothesize that humans and computers working together can create more efficiently and effectively together than separately. The main intended application, currently, is in new product development, particularly in the ideation and concept screening phases.
In no particular order, we hope to accomplish the following with Blender:
- Test our above hypothesis and publish the results in an academic journal, using some version of Blender.
- Improve Blender by making the suggestions more human-friendly.
- Refine our understanding of the intended use-case for Blender. We expect the current version of Blender to benefit domain specialists more than generalists.
- Improve Blender by adding functionality, such as
- An automated domain selection tool. That is, something that would aid users in their choice of seed term, and would likely make use of information about the user's skillset.
- An automated, domain-specific idea evaluator, in order to process a large volume of new ideas quickly. The evaluator would need to attempt to predict the feasibility of the idea, whether something similar exists, and whether the idea is likely to succeed. The current evluation methods are based on a paper by two professors at Columbia.