Jigsaw: Supporting Designers to Prototype Multimodal Applications by Assembling AI Foundation Models
2024

Jigsaw is a prototype system that lets designers generate and alter creative content with AI foundation models represented as puzzle pieces. Designers can search for foundation model capabilities via the Catalog Panel and combine capabilities across different modalities by assembling compatible pieces on the Assembly Panel. Designers can specify inputs and observe intermediate results via the Input and Output Panels. Designers can request the Assembly Assistant to recommend a chain of foundation models to accomplish a specified task.
Recent advancements in AI foundation models have made it possible for them to be utilized off-the-shelf for creative tasks, including ideating design concepts or generating visual prototypes. However, integrating these models into the creative process can be challenging as they often exist as standalone applications tailored to specific tasks. To address this challenge, we introduce Jigsaw, a prototype system that employs puzzle pieces as metaphors to represent foundation models. Jigsaw allows designers to combine different foundation model capabilities by assembling compatible puzzle pieces. To inform the design of Jigsaw, we interviewed ten designers and distilled design goals. In a user study, we showed that Jigsaw enhanced designers’ understanding of available foundation models, provided guidance on combining models across different modalities and capabilities, and served as a canvas to support exploration, prototyping, and design documentation.
Learn more:
Researchers
David Chuan-en Lin, Nikolas Martelaro
Research Areas
Applied Machine Learning, Computational Creativity, Human-Centered AI, Tools