What's the best place for an AI conference, Vancouver or: Why completing comparative questions is difficult

Keyphrases

• Comparative Questions 100%
• Vancouver 100%
• Neural Language Models 66%
• Training Model 33%
• India 33%
• Human Performance 33%
• Training Set 33%
• Publicly Available 33%
• Linguistic Competence 33%
• Text Corpus 33%
• Human Interaction 33%
• Fill-in 33%
• Yield State 33%
• Human Judgment 33%
• Knowledge Skills 33%
• Language Model 33%
• World Knowledge 33%
• Text Model 33%
• General Text 33%
• Large Corpora 33%
• Deep Model 33%
• Evaluation Dataset 33%
• Benchmark Problems 33%
• Competence-based 33%
• NLP Tasks 33%
• Model Worlds 33%

Computer Science

• Artificial Intelligence 100%
• Language Modeling 100%
• Bidirectional Encoder Representations From Transformers 33%
• Subdomains 33%
• World Knowledge 33%
• Benchmark Problem 33%
• Human Interaction 33%