Abstract
Taxonomies provide key information about the semantic relationships between concepts and the inherent organization of vision and language. Despite their impressive capabilities, large multimodal models (LMMs) often lack taxonomic knowledge, leading to low hierarchical visual recognition (HVR) consistency. These models typically only rely on language modeling objectives during fine-tuning and lack explicit taxonomy-aware regularization.
To address this, we propose Hierarchical Representation Regularization ($HiR^2$), a simple plug-and-play regularizer that improves hierarchical consistency in LMMs.
Specifically, we introduce a semantic-aware visual tree construction framework that extracts coarse-to-fine visual features from intermediate LLM layers guided by textual cues. The regularizer combines two complementary objectives: a taxonomic entailment loss that enforces hierarchy via hyperbolic entailment cones in the Lorentz model, and a discriminative dispersive loss that promotes angular separation of semantically similar embeddings on the unit sphere without disturbing the radial hierarchical structure.
Extensive experiments demonstrate that $HiR^2$ effectively captures taxonomic structures across diverse LMMs and fine-tuning methods. Code is available at GitHub.
Blogger's Review: The proposed $HiR^2$ method introduces taxonomic structure regularization into multimodal models, enhancing hierarchical consistency with significant application prospects and research value. The semantic-aware visual tree construction not only strengthens the model's visual understanding but also offers new insights for future model design.