Example use cases

Here we provide example applications of the Seldonian Toolkit, separated by topic. We provide fully runnable code for many examples in Google Colab notebooks. We hope to have frequent additions to this page. At the bottom of the page, we also outline a few use cases of the toolkit that we could envision being used in industry.

Computer vision
  1. Example: Gender bias in facial recognition
Natural Language Processing
  1. Example: Fairness for lie detection
Reinforcement learning
  1. Improving type 1 diabetes treatment
Financial/lending
  1. Fairness for Automated Loan Approval Systems
Student competition winners

Two winners were selected from participants in the first Seldonian Toolkit student competition:

  1. Fairness for Breast Cancer Recurrence Prediction by Derek Lacy.
  2. Fairness in Student Course Completion Based on Student Data by Sahil Yerawar, Pranay Reddy, and Varad Pimpalkhute.
Potential business use cases
  1. Fairness for insurance, mortage and other lending practices that use ML
  2. Fairness for automated hiring and university admittance systems
  3. Guaranteed low false negative rates for medical diagnostic software (such as tumor classification from x-ray images)
  4. Unbiased facial recognition software for criminal identification (e.g., from CCTV screenshots)