Why Robust Natural Language Understanding is a Challenge

TL;DR

This paper discusses the challenges of verifying robustness in natural language understanding models, highlighting the difficulties in applying existing verification methods and proposing a new specification for NLU classification.

Contribution

It introduces a verification specification tailored for NLU classification and analyzes the challenges faced by verifiers in this domain.

Findings

Verifiers struggle to produce positive results despite near-linear separability of data.

Challenges in applying vision-based robustness verification methods to NLU tasks.

Discussion of implications for future robustness verification in NLP.

Abstract

With the proliferation of Deep Machine Learning into real-life applications, a particular property of this technology has been brought to attention: robustness Neural Networks notoriously present low robustness and can be highly sensitive to small input perturbations. Recently, many methods for verifying networks' general properties of robustness have been proposed, but they are mostly applied in Computer Vision. In this paper we propose a Verification specification for Natural Language Understanding classification based on larger regions of interest, and we discuss the challenges of such task. We observe that, although the data is almost linearly separable, the verifier struggles to output positive results and we explain the problems and implications.