Learning Fine-grained Domain Generalization via Hyperbolic State Space Hallucination

2025-04-10

Proceedings of the AAAI Conference on Artificial Intelligence

AI Summary

Fine-grained domain generalization (FGDG) faces a core challenge: fine-grained discriminative patterns are inherently subtle and highly susceptible to cross-domain style shifts—such as illumination and color variations—leading to severe performance degradation when source-trained models are deployed on unseen target domains. To address this, we propose the first dedicated FGDG framework comprising two novel components: (1) a State-Space Hallucination (SSH) module that enhances training data’s stylistic diversity via controllable style extrapolation; and (2) a hyperbolic manifold consistency modeling module that leverages hyperbolic embeddings and distance-aware optimization to capture higher-order semantic relationships among fine-grained categories, thereby improving representation robustness. Evaluated on three standard FGDG benchmarks, our method consistently surpasses all existing state-of-the-art approaches, achieving significant gains in fine-grained classification accuracy on unseen domains.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Adversarial Attacks & Robustness
• Application Domains: Humanities & Computational Social Science

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Fine-grained domain generalization (FGDG) aims to learn a fine-grained representation that can be well generalized to unseen target domains when only trained on the source domain data. Compared with generic domain generalization, FGDG is particularly challenging in that the fine-grained category can be only discerned by some subtle and tiny patterns. Such patterns are particularly fragile under the cross-domain style shifts caused by illumination, color and etc. To push this frontier, this paper presents a novel Hyperbolic State Space Hallucination (HSSH) method. It consists of two key components, namely, state space hallucination (SSH) and hyperbolic manifold consistency (HMC). SSH enriches the style diversity for the state embeddings by firstly extrapolating and then hallucinating the source images. Then, the pre- and post- style hallucinate state embeddings are projected into the hyperbolic manifold. The hyperbolic state space models the high-order statistics, and allows a better discernment of the fine-grained patterns. Finally, the hyperbolic distance is minimized, so that the impact of style variation on fine-grained patterns can be eliminated. Experiments on three FGDG benchmarks demonstrate its state-of-the-art performance.

Problem

• Learning fine-grained representation for unseen domains
• Handling subtle patterns under cross-domain style shifts
• Enhancing generalization via hyperbolic state space modeling

Innovation

• Hyperbolic State Space Hallucination for FGDG
• State space hallucination enriches style diversity
• Hyperbolic manifold minimizes style variation impact