2 papers
cs.CV2026
Adaptive Global and Fine-Grained Perceptual Fusion for MLLM Embeddings Compatible with Hard Negative Amplification
Lexiang Hu, Youze Xue, Dian Li +2
Multimodal embeddings serve as a bridge for aligning vision and language, with the two primary implementations -- CLIP-based and MLLM-based embedding models -- both limited to capt…
cs.CV2025
Improve Multi-Modal Embedding Learning via Explicit Hard Negative Gradient Amplifying
Youze Xue, Dian Li, Gang Liu
With the rapid advancement of multi-modal large language models (MLLMs) in recent years, the foundational Contrastive Language-Image Pretraining (CLIP) framework has been successfu…