Potato Featured at EMNLP 2022

We're proud to announce that our paper introducing Potato was accepted at EMNLP 2022, one of the premier conferences in Natural Language Processing. This milestone represents years of research and development aimed at making data annotation more accessible to the research community.

The Paper

"Potato: The Portable Text Annotation Tool" presents the design philosophy, architecture, and capabilities of Potato. The paper demonstrates how a configuration-first approach can dramatically reduce the barrier to entry for creating high-quality annotation interfaces.

Key Contributions

1. Configuration-First Design: We show how complex annotation interfaces can be specified entirely through YAML configuration, eliminating the need for custom code in most use cases.

2. Flexible Annotation Schemes: Potato supports a wide range of annotation types (radio buttons, checkboxes, spans, Likert scales, and more) that can be combined to create sophisticated annotation tasks.

3. Built-in Quality Control: The paper describes our approach to managing annotator quality through training phases, attention checks, and inter-annotator agreement monitoring.

4. Crowdsourcing Integration: We demonstrate seamless integration with platforms like Prolific and Amazon Mechanical Turk for large-scale annotation studies.

Citing Potato

If you use Potato in your research, please cite our paper:

@inproceedings{pei2022potato,
  title={Potato: The Portable Text Annotation Tool},
  author={Pei, Jiaxin and Anber, Aparna and Jurgens, David},
  booktitle={Proceedings of the 2022 Conference on Empirical Methods
             in Natural Language Processing: System Demonstrations},
  pages={327--337},
  year={2022}
}

Why We Built Potato

The motivation for Potato came from our own frustrations as NLP researchers. Every new annotation project seemed to require either:

1. Learning a complex annotation platform with features we didn't need
2. Building custom interfaces from scratch
3. Compromising on the annotation experience due to tool limitations

We wanted a tool that was:

• Simple: Get started in minutes, not days
• Flexible: Support any annotation task we could imagine
• Portable: Run anywhere without complex infrastructure
• Research-Friendly: Designed for academic workflows and reproducibility

Impact and Adoption

Since its release, Potato has been adopted by research groups worldwide for projects including:

• Sentiment analysis and emotion detection
• Named entity recognition and relation extraction
• Content moderation and toxicity detection
• Argument mining and stance detection
• Clinical NLP and biomedical text mining

Looking Forward

The EMNLP publication was just the beginning. Since then, we've added:

• Image and audio annotation support
• AI-powered annotation assistance
• Active learning integration
• Enhanced collaboration features

We're committed to continuing development based on community feedback. If you have feature requests or ideas, please open an issue on our GitHub repository.

Acknowledgments

We thank the annotators, beta testers, and early adopters who helped shape Potato. Special thanks to the University of Michigan for supporting this research and to the EMNLP reviewers for their valuable feedback.

Read the full paper on ACL Anthology or watch our demo video.