Do Neural Models Learn Systematicity of Monotonicity Inference in Natural Language?

TL;DR

This paper evaluates neural language models' ability to learn systematic monotonicity inference in natural language, revealing they generalize well only when sentence structures are similar to training data.

Contribution

It introduces a novel evaluation method for assessing neural models' systematicity in monotonicity inference and highlights their structural generalization limitations.

Findings

Models generalize to unseen lexical/logical combinations with similar syntax.

Performance drops significantly with slight structural changes in test sentences.

Neural models' generalization is limited to nearly identical syntactic structures.

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.