HELP: A Dataset for Identifying Shortcomings of Neural Models in Monotonicity Reasoning
This work addresses a specific challenge in natural language inference for AI researchers, but it is incremental as it focuses on dataset augmentation rather than novel model architectures.
The authors tackled the problem of neural models' poor performance on monotonicity reasoning in natural language inference by introducing the HELP dataset for data augmentation, which improved overall accuracy, with better gains on lexical replacements than on downward inferences with disjunction and modification.
Large crowdsourced datasets are widely used for training and evaluating neural models on natural language inference (NLI). Despite these efforts, neural models have a hard time capturing logical inferences, including those licensed by phrase replacements, so-called monotonicity reasoning. Since no large dataset has been developed for monotonicity reasoning, it is still unclear whether the main obstacle is the size of datasets or the model architectures themselves. To investigate this issue, we introduce a new dataset, called HELP, for handling entailments with lexical and logical phenomena. We add it to training data for the state-of-the-art neural models and evaluate them on test sets for monotonicity phenomena. The results showed that our data augmentation improved the overall accuracy. We also find that the improvement is better on monotonicity inferences with lexical replacements than on downward inferences with disjunction and modification. This suggests that some types of inferences can be improved by our data augmentation while others are immune to it.