Self-Supervised Document Similarity Ranking via Contextualized Language Models and Hierarchical Inference

Keyphrases

• Document Similarity 100%
• Similarity Score 100%
• Contextual Language Models 100%
• Model Inference 100%
• Hierarchical Inference 100%
• Long Documents 60%
• Token 20%
• Arbitrary Length 20%
• Large Document 20%
• Publicly Available 20%
• Semantic Similarity 20%
• Modern Methods 20%
• Short Documents 20%
• Document Length 20%
• Case Similarity 20%
• Similarity Evaluation 20%
• Longformer 20%
• Self-supervised Method 20%
• Document Datasets 20%

Computer Science

• Language Modeling 100%
• Similarity Ranking 100%
• Supervised Method 25%