Toward Uncertainty Quantification in Modern Art

2026-08-03

AI Summary

Existing uncertainty quantification methods struggle to capture the structural discrepancies among outputs generated with multiple random seeds in modern artistic animation synthesis—such as multimodality, outliers, and fidelity to the original artwork. This work introduces the first structured uncertainty analysis framework for this task, proposing a hybrid evaluation protocol that integrates source-blind and reference-aware assessments. The protocol encompasses 13 distribution estimators, 7-dimensional distribution profiles, and 8 diagnostic question types, supported by the first benchmark corpus of 1,000 videos. Leveraging diverse distribution models—including von Mises–Fisher, Kent, angular central Gaussian, Student-t, kernel density, and mixture models—combined with multi-encoder features, the proposed approach achieves a balanced accuracy of 0.98 (vs. random baseline 0.25) in seed-set topology classification and perfect AUROC of 1.00 in outlier detection (compared to only 0.35 for scalar-based methods), while effectively distinguishing between reference-preserving and reference-missing cases under high uncertainty.

Abstract

Asked to animate the same modern artwork under different random seeds, a text to video model returns visibly different films, one reading per seed. Because modern art is ambiguous by intent, this disagreement is signal, not noise. Yet prevailing uncertainty quantification (UQ) collapses a set of generations to a dispersion scalar that says how much the seeds differ but not how: it cannot tell a compact interpretation from a dominant reading plus an outlier, two competing modes, or diffuse instability, nor whether the set still contains a rendering faithful to the original. We present the first study of the structure of generative uncertainty for modern art animation, and a reusable protocol for identifying source blind multiseed uncertainty: a suite of seven source blind and six reference aware estimators; a distributional profile (robust spread, outlier influence, explicit topology, multimodality, anisotropy, leave one seed influence, reference coverage); a distribution model ablation (vMF, Kent, ACG, Student t, kernel, mixture); eight identification questions; and an artwork level statistical protocol. We build the first corpus: 250 modern artwork captions rendered by Wan2.1 14B under four seeds (1000 videos) across 4 encoders, artworks withheld from generation. As a diagnostic the protocol succeeds: it classifies seed set topology at balanced accuracy 0.98 (chance 0.25), isolates the outlier configuration at AUROC 1.00 where a scalar reaches only 0.35, and splits high uncertainty artworks into reference covering (n=97) and reference missing (n=56) diversity, reliably from three seeds and across encoders.

Problem

• uncertainty quantification
• modern art
• text-to-video generation
• multiseed diversity
• generative models

Innovation

• uncertainty quantification
• generative models
• modern art animation
• multiseed analysis
• distributional profiling