Beyond Linear Bottlenecks: Spline-Based Knowledge Distillation for Culturally Diverse Art Style Classification

2025-07-31

AI Summary

Art style classification faces two key challenges: scarcity of expert annotations and difficulty in modeling nonlinear interactions among stylistic features. To address these, we propose a spline-activated dual-teacher self-supervised framework. It replaces conventional MLPs with Kolmogorov–Arnold Networks (KANs) to explicitly capture complex nonlinear dependencies between global composition and local brushstrokes. A dual-path teacher architecture is designed—one pathway encodes local texture patterns, the other hierarchically represents global stylistic attributes—jointly optimized via knowledge distillation. Evaluated on WikiArt and Pandora18k, our method achieves significant Top-1 accuracy gains over state-of-the-art baselines. Linear probe experiments further demonstrate superior discriminability of the learned style representations. This work establishes a new paradigm for culturally diverse art style understanding—offering both interpretability through spline-based feature decomposition and strong generalization across artistic domains.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Machine Learning: Unsupervised & Self-Supervised Learning
• Natural Language Processing: Sentiment Analysis, Stylistic Analysis, and Argument Mining

Application Category

• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding
• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Art style classification remains a formidable challenge in computational aesthetics due to the scarcity of expertly labeled datasets and the intricate, often nonlinear interplay of stylistic elements. While recent dual-teacher self-supervised frameworks reduce reliance on labeled data, their linear projection layers and localized focus struggle to model global compositional context and complex style-feature interactions. We enhance the dual-teacher knowledge distillation framework to address these limitations by replacing conventional MLP projection and prediction heads with Kolmogorov-Arnold Networks (KANs). Our approach retains complementary guidance from two teacher networks, one emphasizing localized texture and brushstroke patterns, the other capturing broader stylistic hierarchies while leveraging KANs' spline-based activations to model nonlinear feature correlations with mathematical precision. Experiments on WikiArt and Pandora18k demonstrate that our approach outperforms the base dual teacher architecture in Top-1 accuracy. Our findings highlight the importance of KANs in disentangling complex style manifolds, leading to better linear probe accuracy than MLP projections.

Problem

• Addressing nonlinear style-feature interactions in art classification
• Improving global compositional context modeling in style analysis
• Reducing labeled data dependency through enhanced distillation

Innovation

• Replaces MLP with KANs for nonlinear modeling
• Uses dual-teacher networks for complementary guidance
• Leverages spline-based activations for precise feature correlations