A Pragmatics-Centered Evaluation Framework for Natural Language Understanding

TL;DR

This paper introduces PragmEval, a comprehensive benchmark for evaluating pragmatic understanding in NLP models, revealing that current pretraining tasks do not produce truly universal representations.

Contribution

It presents PragmEval, a new benchmark combining 11 pragmatics-focused datasets, and demonstrates its utility in exposing limitations of current models' pragmatic understanding.

Findings

Pretraining on natural language inference does not yield universal representations.

PragmEval effectively evaluates pragmatic understanding in NLP models.

Current benchmarks mainly target semantics, neglecting pragmatics.

Abstract

New models for natural language understanding have recently made an unparalleled amount of progress, which has led some researchers to suggest that the models induce universal text representations. However, current benchmarks are predominantly targeting semantic phenomena; we make the case that pragmatics needs to take center stage in the evaluation of natural language understanding. We introduce PragmEval, a new benchmark for the evaluation of natural language understanding, that unites 11 pragmatics-focused evaluation datasets for English. PragmEval can be used as supplementary training data in a multi-task learning setup, and is publicly available, alongside the code for gathering and preprocessing the datasets. Using our evaluation suite, we show that natural language inference, a widely used pretraining task, does not result in genuinely universal representations, which presents a new challenge for multi-task learning.