Neutral NLI Diversity

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Confidence NLI Diversitymethod0.840
  Best-performing variant: aggregates softmax probability mass rather than binary class counts
• Neutral NLI predictions may capture lexical rather than semantic diversityhypothesis0.822
  Hypothesis proposed to explain Neutral NLI Diversity's high performance on decTest but low on conTest
• What is the role of neutral NLI predictions in measuring semantic diversity?question0.815
  Motivates the creation of Neutral NLI Diversity as an ablation
• NLI Diversity is limited by the scope of the NLI model's training dataclaim0.796
  Limitation acknowledged in discussion section
• Confidence NLI Diversity achieves state-of-the-art performance on measuring semantic diversityclaim0.785
  Main performance claim of the paper
• What is the effect of more generalized NLI training data on NLI Diversity performance?question0.766
  Future work question arising from the finding that Combined model did not outperform MNLI model
• A diverse set of responses for a conversation captures contradictory ways one could respond, measurable by an NLI modelhypothesis0.738
  Core hypothesis motivating the NLI Diversity metric
• 95% confidence intervals overlap between Baseline NLI Diversity, Confidence NLI Diversity, SentBERT, and human judgments on conTestfinding0.736
  Indicates lack of statistically significant differences between top methods