Whispers of Doubt Amidst Echoes of Triumph in NLP Robustness
Authors:
- Ashim GuptaAGAshim Gupta
- Ana MarasovićAMAna Marasović
Preprint
Abstract
Do larger and more performant models resolve NLP's longstanding robustness issues? We investigate this question using over 20 models of different sizes spanning different architectural choices and pretraining objectives. We conduct evaluations using (a) out-of-domain and challenge test sets, (b) behavioral testing with CheckLists, (c) contrast sets, and (d) adversarial inputs. Our analysis reveals that not all out-of-domain tests provide insight into robustness. Evaluating with CheckLists and contrast sets shows significant gaps in model performance; merely scaling models does not make them adequately robust. Finally, we point out that current approaches for adversarial evaluations of models are themselves problematic: they can be easily thwarted, and in their current forms, do not represent a sufficiently deep probe of model robustness. We conclude that not only is the question of robustness in NLP as yet unresolved, but even some of the approaches to measure robustness need to be reassessed.Keywords
All Related Versions
This publication has 18 references indexed in Scilit:
- Cross-Task Generalization via Natural Language Crowdsourcing Instructions2022
- The Dangers of Underclaiming: Reasons for Caution When Reporting How NLP Systems Fail2022
- CrossFit: A Few-shot Learning Challenge for Cross-task Generalization in NLP2021
- Types of Out-of-Distribution Texts and How to Detect Them2021
- What Will it Take to Fix Benchmarking in Natural Language Understanding?2021
- Reevaluating Adversarial Examples in Natural Language2020
- AmbigQA: Answering Ambiguous Open-domain Questions2020
- TextAttack: A Framework for Adversarial Attacks, Data Augmentation, and Adversarial Training in NLP2020
- On the Cross-lingual Transferability of Monolingual Representations2020
- A large annotated corpus for learning natural language inference2015