GeoFlowVLM learns joint distributions of l2-normalized VLM embeddings on the product hypersphere via Riemannian flow matching to expose both aleatoric and epistemic uncertainty through derived entropy and typicality scores.
Mind the gap: Understanding the modality gap in multi-modal contrastive representation learning
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
ToMA uses persistent homology on H0-death and lightweight H1-birth edges to align multimodal manifolds, delivering stable gains on remote sensing and consistent benefits on fashion retrieval.
citing papers explorer
-
GeoFlowVLM: Geometry-Aware Joint Uncertainty for Frozen Vision-Language Embedding
GeoFlowVLM learns joint distributions of l2-normalized VLM embeddings on the product hypersphere via Riemannian flow matching to expose both aleatoric and epistemic uncertainty through derived entropy and typicality scores.
-
Topology-Aware Representation Alignment for Semi-Supervised Vision-Language Learning
ToMA uses persistent homology on H0-death and lightweight H1-birth edges to align multimodal manifolds, delivering stable gains on remote sensing and consistent benefits on fashion retrieval.