airbnb/artificial-adversary

🗣️ Tool to generate adversarial text examples and test machine learning models against them

45
/ 100
Emerging

This tool helps you evaluate and improve the robustness of your automated text classification systems against deliberately disguised messages. It takes your existing text classifier and generates a dataset of modified, 'adversarial' text examples, simulating how users might try to evade detection. The output shows you which types of modifications your classifier is vulnerable to, enabling you to strengthen its performance. This is for data scientists, machine learning engineers, and product managers who deploy text-based machine learning models for tasks like spam detection or content moderation.

402 stars. No commits in the last 6 months.

Use this if you need to understand how well your text classification model performs against intentionally altered text, or if you want to create a more resilient model by training it on diverse, real-world-like adversarial examples.

Not ideal if you are looking for a tool to understand general text sentiment or categorize text into predefined topics without considering deliberate obfuscation.

content-moderation spam-detection fraud-prevention text-analytics abuse-prevention
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 19 / 25

How are scores calculated?

Stars

402

Forks

56

Language

Python

License

MIT

Last pushed

Jan 07, 2022

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/airbnb/artificial-adversary"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.