Stress Test Evaluation for Natural Language Inference
arXiv:1806.00692
Abstract
Natural language inference (NLI) is the task of determining if a natural language hypothesis can be inferred from a given premise in a justifiable manner. NLI was proposed as a benchmark task for natural language understanding. Existing models perform well at standard datasets for NLI, achieving impressive results across different genres of text. However, the extent to which these models understand the semantic content of sentences is unclear. In this work, we propose an evaluation methodology consisting of automatically constructed "stress tests" that allow us to examine whether systems have the ability to make real inferential decisions. Our evaluation of six sentence-encoder models on these stress tests reveals strengths and weaknesses of these models with respect to challenging linguistic phenomena, and suggests important directions for future work in this area.
COLING 2018
Cited by in corpus (6)
- Beyond Leaderboards: A survey of methods for revealing weaknesses in Natural Language Inference data and models
- Several Experiments on Investigating Pretraining and Knowledge-Enhanced Models for Natural Language Inference
- Can NLI Models Verify QA Systems' Predictions?
- SesameBERT: Attention for Anywhere
- Probing Pre-Trained Language Models for Disease Knowledge
- Unsupervised Pre-training with Structured Knowledge for Improving Natural Language Inference