The National Center for Advancing Translational Sciences (NCATS) launched the Bias Detection Tools for Clinical Decision Making Challenge in October 2022. The challenge awarded a combined total of $700,000 to those teams best able to make an open-source solution that detects and mitigates bias in AI/ML models used in healthcare settings. Over 200 individuals registered to compete in this Challenge, and the resources/webinars provided during the course of the Challenge aimed to provide context and cross-training for a complex issue that encompasses AI/ML, healthcare and bias/ethics. The goal of this challenge was to create solutions and share these open-source tools with the community so that the solutions will be enhanced, augmented, integrated, and, optimally, implemented into healthcare systems.
As part of NCATS’ mission to foster sustainable community engagement, sharing, and progress in this space, all outputs (submissions & documentation) are shared back to the community in this repository. Our goal is to further build out our tool-box for making AI/ML unbiased and fair for all subsets of populations; and provide a space for these solutions to allow iteration by stakeholders and to further exhibit and validate the robustness required of these types of tools.
This repository memorializes submissions from the winning teams, honorable mentions, and challengers. The outline below documents each challenger's code and video submission, provides a link back to the original repository, and displays their videos. Additional resources can be found in the resources folder.
First place winner InterFair - Intersectional Fairness with Fairness-Oriented Multiobjective Optimization (FOMO)
Original submission: https://cavalab.org/interfair/
Second place winner iAdeptive AI/ML Bias Detection and Mitigation Tool
Original submission: https://iadeptive.github.io/bias-detection-challenge/
Second place winner MLKFairness
Original submission: https://aryanpoo.github.io/mlkfairness/
Third place winner CVP Bias Detection Tool for Clinical Decision Making
Original submission: https://cvp-bias-detection-healthcare.github.io/
Third place winner BeFair: A Multi-Level-Reweighing Method to Mitigate Bias for Healthcare
Original submission: https://yhzhu99.github.io/BeFair/
Honorable mention GenHealth - AI Bias Detection Informed by Generative AI and FHIR
Original submission: https://genhealth.github.io/
Honorable mention AEquity: A Deep Learning Based Metric for Detecting, Characterizing and Mitigating Dataset Bias
Original submission: https://nadkarni-lab.github.io/ncats-submission-2023/
Honorable mention Metric Lattice for Performance Estimation (MLPE)
Original submission: https://volitico-kellen.github.io/MLPE/MLPE.html
Honorable mention ParaDocs Health: A Free-Text Linguistic Feature Extraction Tool to Detect & Mitigate Provider Implicit Bias Associated with Clinical Decision Making for Hospitalized Patients with Diabetic Ketoacidosis
Original submission: https://voxlej.github.io/ParaDocsHealth_NCATSBiasTool/
Original submission: https://saleh-aldrees.github.io/Bias_Detection_Correction_Tools/
Original submission: https://nicksouligne.github.io/Ziri.github.io/
Original submission: https://github.com/sammyj-w/nih-bias-detect-and-remove-tool/wiki
NIH NCATS Showcase Day - May 5, 2023 from Blue Clarity.
NCATS Bias Detection Webinar 3 2-17-2023 from Blue Clarity.
NCATS Bias Detection Webinar 2 1-30-2023 from Blue Clarity.
