Making OSS Machine Learning communities more diverse
summary
Argues that counting only repo commits and platform metadata produces a biased picture, cites research showing surveys and pull‑request metrics exclude documentation, dataset curation, design and governance work, and pushes Data Feminism Principle #7 to make labor visible. Recommends adopting a measurement framework such as Preventing Random Acts of Metrics, tracking dataset and documentation contributions, using programs like Google Summer of Code for onboarding, and points to CPython and TensorFlow as projects actively improving diversity.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 0:19Why is diversity important
- 2:53What gets counted
- 4:18Find your open source
- 5:42QA
- 5:56Google Summer of Code
- 6:40Measuring OSS contributions
- 8:32How do I start
- 10:37How do I showcase
- 13:33How to prioritize
related talks
30mTokenism in Open Source: Diversity or Decoration?
39mState of diversity in open source communities
23mKeeping Your Open Source Project Accessible to All
23mEquity and inclusion as a way of developing diverse, resilient, and more relevant OSS
36mPanel Discussion: Metrics that Matter: How to Forge More Diverse, Inclusive Communities
25mThis is bigger than us: Building a future for Open Source
31mHow to amplify diversity inside a Python community?
32mBuilding inclusive Open Source communities