Visual Perturbation and Adaptive Hard Negative Contrastive Learning for Compositional Reasoning in Vision-Language Models

Ähnliche Einträge

• No Hard Negatives Required: Concept Centric Learning Leads to Compositionality without Degrading Zero-shot Capabilities of Contrastive Models
  von: Pham, Hai X., et al.
  Veröffentlicht: (2026)
• Unleashing the Power of Vision-Language Models for Long-Tailed Multi-Label Visual Recognition
  von: Tang, Wei, et al.
  Veröffentlicht: (2025)
• Qwen Look Again: Guiding Vision-Language Reasoning Models to Re-attention Visual Information
  von: Chu, Xu, et al.
  Veröffentlicht: (2025)
• Semantic Compositions Enhance Vision-Language Contrastive Learning
  von: Aladago, Maxwell, et al.
  Veröffentlicht: (2024)
• Negative Label Guided OOD Detection with Pretrained Vision-Language Models
  von: Jiang, Xue, et al.
  Veröffentlicht: (2024)