Artographer: a Curatorial Interface for Art Space Exploration

2025-12-02

AI Summary

Contemporary AI-powered art interfaces often obscure, rather than reveal, latent semantic and contextual relationships among artworks. Method: We propose a “curatorial interface” framework that constructs a scalable, two-dimensional semantic map of over 15,000 historical artworks using embeddings from pretrained vision models, combined with clustering and dimensionality reduction for interpretable visualization. Users explore the map via spatial interactions—including jumping, roaming, focusing, and revisiting—while behavioral data from 20 participants (including 9 art historians) are analyzed to evaluate interaction patterns. Contribution/Results: Our evaluation demonstrates that spatial mapping significantly improves both the efficiency of relationship discovery and the depth of exploratory engagement. It further reveals how core curatorial values—visibility, serendipity, friction, and user autonomy—are embodied in interaction modalities. The work establishes a novel design paradigm for AI-augmented art curation systems that meaningfully integrate computational capability with humanistic agency and interpretive subjectivity.

Technology Category

• Humans and AI: Intelligent User Interfaces
• Cognitive Modeling & Cognitive Systems: Computational Creativity
• Computer Vision: Interpretability, Explainability, and Transparency

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Relating a piece to previously established works is crucial in creating and engaging with art, but AI interfaces tend to obscure such relationships, rather than helping users explore them. Embedding models present new opportunities to support discovering and relating artwork through spatial interaction. We built Artographer, an art exploration system featuring a zoomable 2-D map, constructed from the similarity-clustered embeddings of 15,000+ historical artworks. Using Artographer as a probe to investigate spatial artwork exploration, we analyzed how 20 participants (including 9 art history scholars) traversed the map, during a goal-driven task and when freely exploring. We observe divergent and convergent exploration behaviors (Jumping, Wandering, Fixation, Revisiting) and identify values enacted by spatial art-finding (Visibility, Agency, Serendipity, Friction.) We situate spatial maps within a space of Curatorial Interfaces, systems that select and present artworks, and discuss centering pluralism and agency in the design of more responsible AI systems for art curation.

Problem

• Develops a spatial interface for exploring relationships among historical artworks.
• Investigates user behaviors and values in AI-driven art exploration systems.
• Proposes curatorial interfaces emphasizing pluralism and agency in AI design.

Innovation

• Embedding models enable spatial art exploration
• Zoomable 2-D map clusters 15,000+ artwork embeddings
• Spatial interface supports divergent and convergent behaviors