Efficient Visual Appearance Optimization by Learning from Prior Preferences

2025-07-21

AI Summary

To address the high interaction cost and insufficient personalization in visual parameter tuning, this paper proposes Meta-PO—the first method integrating meta-learning into the Preference-based Bayesian Optimization (PBO) framework. Meta-PO implicitly models cross-user preference priors to enable rapid cold-start adaptation and user-specific customization; it further incorporates an intelligent candidate design generation mechanism to substantially improve query efficiency. Experiments demonstrate that, under similar visual objectives, users achieve satisfactory results in only 5.86 interactions on average; under diverse objectives, tuning converges within eight interactions—significantly reducing interaction overhead compared to conventional PBO approaches. This work establishes the first systematic paradigm unifying meta-learning with preference optimization, offering a novel pathway toward efficient, low-burden visual customization for end users.

Technology Category

• Search and Optimization: Metareasoning and Metaheuristics
• Machine Learning: Learning Preferences or Rankings
• Computer Vision: Learning & Optimization for CV

Application Category

• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Fairness, privacy, and diversity in economic environments

Abstract

Adjusting visual parameters such as brightness and contrast is common in our everyday experiences. Finding the optimal parameter setting is challenging due to the large search space and the lack of an explicit objective function, leaving users to rely solely on their implicit preferences. Prior work has explored Preferential Bayesian Optimization (PBO) to address this challenge, involving users to iteratively select preferred designs from candidate sets. However, PBO often requires many rounds of preference comparisons, making it more suitable for designers than everyday end-users. We propose Meta-PO, a novel method that integrates PBO with meta-learning to improve sample efficiency. Specifically, Meta-PO infers prior users' preferences and stores them as models, which are leveraged to intelligently suggest design candidates for the new users, enabling faster convergence and more personalized results. An experimental evaluation of our method for appearance design tasks on 2D and 3D content showed that participants achieved satisfactory appearance in 5.86 iterations using Meta-PO when participants shared similar goals with a population (e.g., tuning for a ``warm'' look) and in 8 iterations even generalizes across divergent goals (e.g., from ``vintage'', ``warm'', to ``holiday''). Meta-PO makes personalized visual optimization more applicable to end-users through a generalizable, more efficient optimization conditioned on preferences, with the potential to scale interface personalization more broadly.

Problem

• Optimizing visual parameters without explicit objectives
• Reducing preference comparison rounds for end-users
• Personalizing design suggestions using prior user preferences

Innovation

• Integrates PBO with meta-learning for efficiency
• Infers prior preferences to suggest designs
• Achieves faster convergence with personalized results