Scalable attentive sentence-pair modeling via distilled sentence embedding

Keyphrases

• Sentence Embedding 100%
• Sentence Pair Modeling 100%
• Embedding Method 25%
• Sentence Pair 25%
• Cross-attention 25%
• Teacher Model 25%
• Order of Magnitude 12%
• Art Performance 12%
• Natural Language Understanding 12%
• XLNet 12%
• Elementary Operation 12%
• Attentive Model 12%
• Sentence Representation 12%
• Fine-tuned BERT 12%
• Vector Mapping 12%
• Attention Layer 12%
• Student Model 12%
• Knowledge Distillation 12%

Computer Science

• Bidirectional Encoder Representations From Transformers 100%
• Embedding Method 33%
• Art Performance 33%
• Natural-Language Understanding 33%
• Universal Sentence 33%
• Knowledge Distillation 33%