Do Neural Models Learn Systematicity of Monotonicity Inference in Natural Language?
arXiv:2004.14839
Abstract
Despite the success of language models using neural networks, it remains unclear to what extent neural models have the generalization ability to perform inferences. In this paper, we introduce a method for evaluating whether neural models can learn systematicity of monotonicity inference in natural language, namely, the regularity for performing arbitrary inferences with generalization on composition. We consider four aspects of monotonicity inferences and test whether the models can systematically interpret lexical and logical phenomena on different training/test splits. A series of experiments show that three neural models systematically draw inferences on unseen combinations of lexical and logical phenomena when the syntactic structures of the sentences are similar between the training and test sets. However, the performance of the models significantly decreases when the structures are slightly changed in the test set while retaining all vocabularies and constituents already appearing in the training set. This indicates that the generalization ability of neural models is limited to cases where the syntactic structures are nearly the same as those in the training set.
arXiv commentaccepted by ACL2020 as a long paper
References in corpus (4)
- Assessing BERT's Syntactic Abilities
- Performance Impact Caused by Hidden Bias of Training Data for Recognizing Textual Entailment
- Siamese recurrent networks learn first-order logic reasoning and exhibit zero-shot compositional generalization
- Compositionality decomposed: how do neural networks generalise?