Learning to Prompt for Vision-Language Models

Keyphrases

• Vision-language Models 100%
• Image Recognition 66%
• Downstream Task 66%
• Prompt Engineering 66%
• Recent Advances 33%
• Natural Language Processing 33%
• Recognition Task 33%
• Generalization Performance 33%
• Optimization Model 33%
• Learning-based 33%
• Learning Representations 33%
• Representation Learning 33%
• Domain Generalization 33%
• Discretized 33%
• Learning Research 33%
• Impact Performance 33%
• Class-specific 33%
• Natural Language 33%
• Classifier Weights 33%
• Pre-Trained Parameters 33%
• Over 45 33%
• Domain Expertise 33%
• Prompt Learning 33%
• Context Word 33%
• Zero-shot Models 33%
• Common Feature Space 33%
• Zero-shot Transfer 33%
• Average Gain 33%
• Vision-language Pre-training 33%

Computer Science

• Language Modeling 100%
• Representation Learning 66%
• Prompt Engineering 66%
• Zero-Shot Learning 66%
• Natural Language Processing 33%
• Generalization Performance 33%
• Feature Space 33%
• Domain Expertise 33%
• Requires Optimization 33%
• Prompt Learning 33%