An Investigation of the (In)effectiveness of Counterfactually Augmented Data

07/01/2021
by   Nitish Joshi, et al.
0

While pretrained language models achieve excellent performance on natural language understanding benchmarks, they tend to rely on spurious correlations and generalize poorly to out-of-distribution (OOD) data. Recent work has explored using counterfactually-augmented data (CAD) – data generated by minimally perturbing examples to flip the ground-truth label – to identify robust features that are invariant under distribution shift. However, empirical results using CAD for OOD generalization have been mixed. To explain this discrepancy, we draw insights from a linear Gaussian model and demonstrate the pitfalls of CAD. Specifically, we show that (a) while CAD is effective at identifying robust features, it may prevent the model from learning unperturbed robust features, and (b) CAD may exacerbate existing spurious correlations in the data. Our results show that the lack of perturbation diversity in current CAD datasets limits its effectiveness on OOD generalization, calling for innovative crowdsourcing procedures to elicit diverse perturbation of examples.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
02/18/2023

Improving the Out-Of-Distribution Generalization Capability of Language Models: Counterfactually-Augmented Data is not Enough

Counterfactually-Augmented Data (CAD) has the potential to improve langu...
research
09/14/2021

How Does Counterfactually Augmented Data Impact Models for Social Computing Constructs?

As NLP models are increasingly deployed in socially situated settings su...
research
09/17/2023

Mitigating Shortcuts in Language Models with Soft Label Encoding

Recent research has shown that large language models rely on spurious co...
research
05/09/2022

Counterfactually Augmented Data and Unintended Bias: The Case of Sexism and Hate Speech Detection

Counterfactually Augmented Data (CAD) aims to improve out-of-domain gene...
research
11/29/2022

AutoCAD: Automatically Generating Counterfactuals for Mitigating Shortcut Learning

Recent studies have shown the impressive efficacy of counterfactually au...
research
05/24/2023

L-CAD: Language-based Colorization with Any-level Descriptions

Language-based colorization produces plausible and visually pleasing col...
research
03/09/2023

Optimizing CAD Models with Latent Space Manipulation

When it comes to the optimization of CAD models in the automation domain...

Please sign up or login with your details

Forgot password? Click here to reset