Aggregate-and-Adapt Natural Language Prompts for Downstream Generalization of CLIP
–Neural Information Processing Systems
Large pretrained vision-language models like CLIP have shown promising generalization capability, but may struggle in specialized domains ( e.g., satellite imagery)
Neural Information Processing Systems
Feb-16-2026, 17:35:29 GMT
- Genre:
- Research Report > Experimental Study (0.93)
- Technology:
- Information Technology > Artificial Intelligence
- Machine Learning > Neural Networks (0.68)
- Natural Language
- Large Language Model (0.73)
- Text Processing (0.46)
- Representation & Reasoning (1.00)
- Vision (1.00)
- Information Technology > Artificial Intelligence