soarsmu/BiasFinder
BiasFinder | IEEE TSE | Metamorphic Test Generation to Uncover Bias for Sentiment Analysis Systems
This project helps sentiment analysis practitioners identify hidden biases in their AI systems. It takes an existing sentiment analysis model and text data, then generates new texts that subtly change demographic details like gender or occupation. The output highlights specific text pairs where the system unexpectedly changes its sentiment prediction, revealing potential bias against certain groups. This is for data scientists, machine learning engineers, or AI ethicists who deploy or manage sentiment analysis systems.
No commits in the last 6 months.
Use this if you need to systematically test your sentiment analysis system for demographic bias related to gender, occupation, or country of origin.
Not ideal if you are looking for a general-purpose bias detection tool for other AI systems or data types beyond text sentiment analysis.
Stars
11
Forks
6
Language
Jupyter Notebook
License
Apache-2.0
Category
Last pushed
Jan 18, 2022
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/nlp/soarsmu/BiasFinder"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
dccuchile/wefe
WEFE: The Word Embeddings Fairness Evaluation Framework. WEFE is a framework that standardizes...
dreji18/Fairness-in-AI
Detecting Bias and ensuring Fairness in AI solutions
amazon-science/bold
Dataset associated with "BOLD: Dataset and Metrics for Measuring Biases in Open-Ended Language...
dhfbk/variationist
Variationist: Exploring Multifaceted Variation and Bias in Written Language Data (ACL 2024 demo track)
microsoft/SafeNLP
Safety Score for Pre-Trained Language Models