May AI?

Abstract

Design ideation is a prime creative activity in design. However, it is challenging to support computationally due to its quickly evolving and exploratory nature. The paper presents cooperative contextual bandits (CCB) as a machine-learning method for interactive ideation support. A CCB can learn to propose domain-relevant contributions and adapt their exploration/exploitation strategy. We developed a CCB for an interactive design ideation tool that 1) suggests inspirational and situationally relevant materials ("may AI?"); 2) explores and exploits inspirational materials with the designer; and 3) explains its suggestions to aid reflection. The application case of digital mood board design is presented, wherein visual inspirational materials are collected and curated in collages. In a controlled study, 14 of 16 professional designers preferred the CCB-augmented tool. The CCB approach holds promise for ideation activities wherein adaptive and steerable support is welcome but designers must retain full outcome control. CCS CONCEPTS • Human-centered computing → Interactive systems and tools; • Computing methodologies → Machine learning; • Applied computing → Arts and humanities.

Self-supervised Multi-view Disentanglement for Expansion of Visual Collections

“…We also intend to further evaluate the benefits of our approach via a thorough user study. From the perspective of the application that we have considered, extending our current setup to multiple iterations [13,29] is a natural next step. While the treatment in the current paper is restricted to specific visual axes or views, the proposed framework is generalizable to a broader range of visual representation spaces.…”

“…Facilitating design ideation via moodboards has been explored by Rieuf et al [23], though their focus was an immersive interface into the corpus of assets. The recent work of Koch et al [13] recommends that the exploratory process of ideation be an interactive one, with the system each time refining its view of the user's intent. Solving for the needs of creatives involves a holistic treatment that includes interface, interaction and many more dimensions.…”

Composing Mood Board with User Feedback in Concept Space

“…Koch et al [10] focused on the problems in initial design ideation with an exploration-exploitation trade-off, which is the balance between adhering to the option that yielded the highest benefit in the past and exploring new options that might offer higher payoffs in the future [11]. They created an interactive digital tool to support designers in creating a mood board, utilizing this exploration-exploitation strategy optimized by a cooperative contextual bandit reinforcement learning algorithm.…”

Learning Contextualized User Preferences for Co‐Adaptive Guidance in Mixed‐Initiative Topic Model Refinement

“…While these approaches provide guidance upon request, we propose a technique in which agents learn to provide guidance automatically when appropriate. This paper was loosely inspired by cooperative contextual bandits [TvdS15], a class of recommender systems that have recently been used successfully in personalized design ideation for the creation of mood boards [KLHO19]. However, this approach relies on partitioning the design space of mood boards according to multiple dimensions and assigning agents to each partition that then sample suggestions from their partition.…”