Efficient Personalization of Generative User Interfaces

2026-04-10

AI Summary

This work addresses the challenge of personalizing generative user interfaces, where ideal interface attributes are highly subjective, difficult to articulate explicitly, and costly to learn from sparse user feedback. The authors construct a dataset comprising pairwise preferences from 20 designers over 600 generated interfaces, revealing substantial disagreement in design preferences. They propose an efficient personalization approach that models a new user’s preferences as a linear combination of existing designers’ preferences, eschewing predefined design rules. Combining lightweight contrastive preference learning with generative optimization, their method outperforms both pretrained UI evaluators and large multimodal models in technical evaluations. In a user study with 12 new designers, the generated interfaces significantly surpassed baseline approaches such as direct prompting, demonstrating the method’s effectiveness and practical utility.

Technology Category

• Humans and AI: Intelligent User Interfaces
• Machine Learning: Learning Preferences or Rankings
• Knowledge Representation and Reasoning: Preferences

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications

Abstract

Generative user interfaces (UIs) create new opportunities to adapt interfaces to individual users on demand, but personalization remains difficult because desirable UI properties are subjective, hard to articulate, and costly to infer from sparse feedback. We study this problem through a new dataset in which 20 trained designers each provide pairwise judgments over the same 600 generated UIs, enabling direct analysis of preference divergence. We find substantial disagreement across designers (average kappa = 0.25), and written rationales reveal that even when designers appeal to similar concepts such as hierarchy or cleanliness, designers differ in how they define, prioritize, and apply those concepts. Motivated by these findings, we develop a sample-efficient personalization method that represents a new user in terms of prior designers rather than a fixed rubric of design concepts. In a technical evaluation, our preference model outperforms both a pretrained UI evaluator and a larger multimodal model, and scales better with additional feedback. When used to personalize generation, it also produces interfaces preferred by 12 new designers over baseline approaches, including direct user prompting. Our findings suggest that lightweight preference elicitation can serve as a practical foundation for personalized generative UI systems.

Problem

• personalization
• generative user interfaces
• subjective preferences
• preference elicitation
• UI design

Innovation

• generative user interfaces
• personalization
• preference modeling
• sample-efficient learning
• designer disagreement