On the Efficacy of Adversarial Data Collection for Question Answering: Results from a Large-Scale Randomized Study

06/02/2021
by   Divyansh Kaushik, et al.
0

In adversarial data collection (ADC), a human workforce interacts with a model in real time, attempting to produce examples that elicit incorrect predictions. Researchers hope that models trained on these more challenging datasets will rely less on superficial patterns, and thus be less brittle. However, despite ADC's intuitive appeal, it remains unclear when training on adversarial datasets produces more robust models. In this paper, we conduct a large-scale controlled study focused on question answering, assigning workers at random to compose questions either (i) adversarially (with a model in the loop); or (ii) in the standard fashion (without a model). Across a variety of models and datasets, we find that models trained on adversarial data usually perform better on other adversarial datasets but worse on a diverse collection of out-of-domain evaluation sets. Finally, we provide a qualitative analysis of adversarial (vs standard) data, identifying key differences and offering guidance for future research.

READ FULL TEXT

page 1

page 2

page 3

page 4

12/16/2021

Models in the Loop: Aiding Crowdworkers with Generative Annotation Assistants

In Dynamic Adversarial Data Collection (DADC), human annotators are task...
04/18/2021

Improving Question Answering Model Robustness with Synthetic Adversarial Data Generation

Despite the availability of very large datasets and pretrained models, s...
01/31/2022

Adaptive Sampling Strategies to Construct Equitable Training Datasets

In domains ranging from computer vision to natural language processing, ...
06/04/2021

Human-Adversarial Visual Question Answering

Performance on the most commonly used Visual Question Answering dataset ...
11/16/2021

Adversarially Constructed Evaluation Sets Are More Challenging, but May Not Be Fair

More capable language models increasingly saturate existing task benchma...
09/27/2018

A Qualitative Comparison of CoQA, SQuAD 2.0 and QuAC

In this work, we compare three datasets which build on the paradigm defi...