A large annotated corpus for learning natural language inference

1 Introduction

1. What is natural language inference (NLI), and why is it a good test of semantic representations?
   Deciding whether a hypothesis sentence is entailed by, contradicts, or is neutral to (independent of) a premise. Entailment and contradiction are central to meaning, from the lexicon to whole texts, so NLI is a testing ground for representations that must support logical and commonsense inference.
   • Bowman et al., 2015, A large annotated corpus for learning natural language inference (Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing (EMNLP)), §1, pp. 632–633
2. What problems with existing NLI corpora motivated SNLI?
   They were too small for data-hungry models (RTE sets under a thousand examples each, SICK 4,500 training pairs), many sentences were algorithmically generated (introducing spurious patterns), and labels suffered from indeterminacies of event and entity coreference.
   • Bowman et al., 2015, A large annotated corpus for learning natural language inference (Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing (EMNLP)), §1, pp. 632–633
3. What is SNLI, in numbers?
   The Stanford Natural Language Inference corpus: 570,152 sentence pairs labeled entailment, contradiction or neutral, two orders of magnitude larger than other resources of its kind, with all sentences and labels written by humans in a grounded, naturalistic task.
   • Bowman et al., 2015, A large annotated corpus for learning natural language inference (Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing (EMNLP)), §1, pp. 632–633