Exploring Transitivity in Neural NLI Models through Veridicality

TL;DR

This paper investigates whether neural models for natural language inference can perform transitive reasoning, revealing that current models struggle with compositional inference tasks involving clause-embedding verbs.

Contribution

It introduces a novel analysis method using synthetic and naturalistic datasets to evaluate transitivity inferences in NLI models, highlighting their limitations.

Findings

Current NLI models perform poorly on transitivity inference tasks.

Models lack the ability to generalize compositional inference patterns.

Analysis code and datasets are publicly available.

Abstract

Despite the recent success of deep neural networks in natural language processing, the extent to which they can demonstrate human-like generalization capacities for natural language understanding remains unclear. We explore this issue in the domain of natural language inference (NLI), focusing on the transitivity of inference relations, a fundamental property for systematically drawing inferences. A model capturing transitivity can compose basic inference patterns and draw new inferences. We introduce an analysis method using synthetic and naturalistic NLI datasets involving clause-embedding verbs to evaluate whether models can perform transitivity inferences composed of veridical inferences and arbitrary inference types. We find that current NLI models do not perform consistently well on transitivity inference tasks, suggesting that they lack the generalization capacity for drawing composite inferences from provided training examples. The data and code for our analysis are publicly available at https://github.com/verypluming/transitivity.