DQI: Measuring Data Quality in NLP Article Swipe

MAKE IMPACT

Swaroop Mishra , Anjana Arunkumar , Bhavdeep Sachdeva , Chris Bryan , Chitta Baral · YOU? · · 2020 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2005.00816

Neural language models have achieved human level performance across several NLP datasets. However, recent studies have shown that these models are not truly learning the desired task; rather, their high performance is attributed to overfitting using spurious biases, which suggests that the capabilities of AI systems have been over-estimated. We introduce a generic formula for Data Quality Index (DQI) to help dataset creators create datasets free of such unwanted biases. We evaluate this formula using a recently proposed approach for adversarial filtering, AFLite. We propose a new data creation paradigm using DQI to create higher quality data. The data creation paradigm consists of several data visualizations to help data creators (i) understand the quality of data and (ii) visualize the impact of the created data instance on the overall quality. It also has a couple of automation methods to (i) assist data creators and (ii) make the model more robust to adversarial attacks. We use DQI along with these automation methods to renovate biased examples in SNLI. We show that models trained on the renovated SNLI dataset generalize better to out of distribution tasks. Renovation results in reduced model performance, exposing a large gap with respect to human performance. DQI systematically helps in creating harder benchmarks using active learning. Our work takes the process of dynamic dataset creation forward, wherein datasets evolve together with the evolving state of the art, therefore serving as a means of benchmarking the true progress of AI.

Type

preprint

Language

en

Landing Page

http://arxiv.org/abs/2005.00816

PDF

https://arxiv.org/pdf/2005.00816

OA Status

green

Cited By

7

References

76

Related Works

20

OpenAlex ID

https://openalex.org/W3021112832

Raw OpenAlex JSON

OpenAlex ID

https://openalex.org/W3021112832

Canonical identifier for this work in OpenAlex

DOI

https://doi.org/10.48550/arxiv.2005.00816

Digital Object Identifier

Title

DQI: Measuring Data Quality in NLP

Work title

Type

preprint

OpenAlex work type

Language

en

Primary language

Publication year

2020

Year of publication

Publication date

2020-05-02

Full publication date if available

Authors

Swaroop Mishra, Anjana Arunkumar, Bhavdeep Sachdeva, Chris Bryan, Chitta Baral

List of authors in order

Landing page

https://arxiv.org/abs/2005.00816

Publisher landing page

PDF URL

https://arxiv.org/pdf/2005.00816

Direct link to full text PDF

Open access

Yes

Whether a free full text is available

OA status

green

Open access status per OpenAlex

OA URL

https://arxiv.org/pdf/2005.00816

Direct OA link when available

Concepts

Overfitting, Computer science, Benchmarking, Artificial intelligence, Machine learning, Deep learning, Quality (philosophy), Task (project management), Automation, Spurious relationship, Process (computing), Data quality, Data mining, Artificial neural network, Engineering, Business, Metric (unit), Mechanical engineering, Operating system, Philosophy, Epistemology, Systems engineering, Marketing, Operations management

Top concepts (fields/topics) attached by OpenAlex

Cited by

7

Total citation count in OpenAlex

Citations by year (recent)

2021: 2, 2020: 5

Per-year citation counts (last 5 years)

References (count)

76

Number of works referenced by this work

Related works (count)

20

Other works algorithmically related by OpenAlex

Full payload