https://github.com/mariajmolina/innovator_ai_bias
Software associated with the NSF NCAR Early Career Faculty Innovator Program for "Pathways to Mitigating Bias for Environmental Justice in Black Communities"
https://github.com/mariajmolina/innovator_ai_bias
ai bias justice
Last synced: about 2 months ago
JSON representation
Software associated with the NSF NCAR Early Career Faculty Innovator Program for "Pathways to Mitigating Bias for Environmental Justice in Black Communities"
- Host: GitHub
- URL: https://github.com/mariajmolina/innovator_ai_bias
- Owner: mariajmolina
- Created: 2022-07-20T23:12:25.000Z (almost 3 years ago)
- Default Branch: main
- Last Pushed: 2022-11-27T20:21:31.000Z (over 2 years ago)
- Last Synced: 2023-03-10T06:36:08.355Z (about 2 years ago)
- Topics: ai, bias, justice
- Language: Jupyter Notebook
- Homepage:
- Size: 8.58 MB
- Stars: 1
- Watchers: 1
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Innovator_AI_Bias
Repository associated with the NSF-funded, [NCAR Early Career Faculty Innovator Program](https://edec.ucar.edu/university-partnerships/innovators).
Project is being led by Prof. Amy Quarkume from Howard University (PI) with graduate students Mikah Jones and Jessica Moulite.
NCAR scientist collaborators include Julie Demuth, Curtis Walker, Forrest Lacey, and Maria J. Molina.
## Pathways to Mitigating Bias for Environmental Justice in Black Communities
__Overview of Research Project__
This [project](https://edec.ucar.edu/university-partnerships/innovators/2021-2023-cohort) is a three-tier interdisciplinary undertaking in consultation with colleagues at NCAR to understand the impacts of AI Bias, as it relates to Black, Latino/a, and Tribal communities. The main question explores, how can we detect and quantify the extent of societal, cultural, and historical bias in earth science and AI data, which consequently disproportionally impact margined communities? Through survey sampling, mapping, and AI data collection the study will explore the quality of earth science data, output and feedback loops of AI and ML models in five states (Louisiana, Michigan, Mississippi, Texas, and the District of Columbia) representing a cross section of African American, Latino/a and Indigenous Native American communities and cities in the US. Through these methods, the researchers hope to explore how algorithms exacerbate historical disparities for marginalized communities around pollution, weather, transportation, and disaster preparedness.
__Broader Outcomes__
As a social scientist and historian, at the core of this project is the goal of closing the gap between data technology and cultural studies, with the hope of (1) expanding the interdisciplinary field of Africana Studies, AI, and Atmospheric Science as a vehicle to advance Actionable Earth System Science; (2) researching the depth of data bias in pollution, weather forecasting and disaster preparedness in the high population Black, Latino/a and Indigenous Native American communities; and (3) training more students to become culturally conscious scientists to mitigate bias.