One of these words is not like the other: a reproduction of outlier identification using non-contextual word representations

Keyphrases

• Word Embedding 100%
• Outlier Identification 100%
• Contextual Word Representations 100%
• Hyperparameters 100%
• High Performance 50%
• Multiple Languages 50%
• Token 50%
• Neural Model 50%
• Computationally Expensive 50%
• Downstream Task 50%
• Ambiguous Words 50%
• Italian Language 50%
• German Language 50%
• Intrinsic Evaluation 50%
• Cost-aware 50%
• Reproduction Study 50%
• Identification Task 50%
• Multi-keyword 50%
• Greatest Difficulty 50%
• NLP. 50%

Computer Science

• Word Embedding 100%
• Research Community 50%
• Multiple Language 50%
• Syntactics 50%
• Good Performance 50%