Learning User Preferences for Image Generation Model
This work addresses the problem of personalized image generation for users by improving preference prediction, though it appears incremental as it builds on existing multimodal models with specific enhancements.
The paper tackles the problem of predicting user preferences for image generation by addressing the limitations of existing methods that neglect individual variability and dynamic tastes, proposing an approach using Multimodal Large Language Models with contrastive preference loss and preference tokens, which outperforms other methods in preference prediction accuracy.
User preference prediction requires a comprehensive and accurate understanding of individual tastes. This includes both surface-level attributes, such as color and style, and deeper content-related aspects, such as themes and composition. However, existing methods typically rely on general human preferences or assume static user profiles, often neglecting individual variability and the dynamic, multifaceted nature of personal taste. To address these limitations, we propose an approach built upon Multimodal Large Language Models, introducing contrastive preference loss and preference tokens to learn personalized user preferences from historical interactions. The contrastive preference loss is designed to effectively distinguish between user ''likes'' and ''dislikes'', while the learnable preference tokens capture shared interest representations among existing users, enabling the model to activate group-specific preferences and enhance consistency across similar users. Extensive experiments demonstrate our model outperforms other methods in preference prediction accuracy, effectively identifying users with similar aesthetic inclinations and providing more precise guidance for generating images that align with individual tastes. The project page is \texttt{https://learn-user-pref.github.io/}.