Collaborative Ideation Partner (CIP) is a co-creative design system that provides inspirational sketches based on the visual and conceptual similarity to sketches drawn by a designer. The AI models measure similarity in the CIP use deep learning models and cosine similarity to the user’s sketch or design concept. The interactive experience allows the user to seek inspiration as needed. The focus of this study is on how the visual and conceptual similarity of the contribution of the AI partner influences design ideation in a co-creative system.
The AI model for visual similarity selects sketches from the sketch dataset that share some structural characteristics with the user’s sketch. The AI model for conceptual similarity computes the degree of similarity between the category name of the object that the user is designing and the category names in the sketch dataset.
For the source of inspiring sketches, CIP uses a public benchmark dataset called QuickDraw!, which was created during an online game where players were asked to draw a particular object within 20 seconds. The dataset includes 345 categories with more than 50 million labeled sketches, where sketches are the array of the x and y coordinates of the strokes. The stroke data associated with these sketches are used to calculate the visual similarity and the corresponding category names are used to measure the conceptual similarity. For measuring conceptual similarity, we use sketch category names in the QuickDraw dataset as the concepts of the sketches that contain 345 unique categories. We use two word2vec models and calculate cosine similarities for measuring the conceptual similarities between target designs for participants and inspiring sketches from the dataset. The two word2vec models are pretrained on Google News and Wikipedia dataset.
In order to measure design ideation in a co-creative system, we developed an approach for measuring ideation that has two components: an outcome-based approach and a process-based approach. The outcome-based approach adapts existing quantitative metrics for ideation: novelty, variety, quality, and quantity of ideas expressed in the outcome. The process-based approach uses existing cognitive models of design, including the Function-Behavior-Structure (FBS) ontology and the Problem-Solution (P-S) index, to code and analyze the verbal protocol of the designers. These measures can be used in evaluating the impact of AI contributions in other co-creative systems that support design creativity.