Building Inclusive Machine Learning Models
summary
Case studies name concrete harms: the Apple credit card shows a 20x gap in limits, a health‑care algorithm routes people of color to $1,800 less spending on average, and Amazon’s hiring model reproduces historical gender skew. Practical fixes include balanced sampling, fairness constraints like equal opportunity or demographic parity, post‑training corrections and continuous audits, and tools such as Google AutoML for fairness‑aware training. Policy notes cite the EU AI Act’s dataset quality checks and stricter rules for high‑risk systems.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 0:35Who is Elizabeth
- 1:09AI for everyone
- 2:30Apple Credit Card Case Study
- 3:25Health Care Case Study
- 5:49Cancer Detection
- 6:07Bias in Data
- 9:19Key Questions
- 11:15What is Bias
- 14:57Sources Consequences of Bias
- 16:41Fairness in AI
- 19:10Limitations of fairness metrics
- 24:46Holistic approach
- 25:40Mitigation strategies
- 27:05Best practices
- 32:01Policy and global trends
related talks
23mMachine Learning Ethics
21mFairness and bias in AI
37mKeynote: Ethical Machine Learning: Creating Fair Models in an Unjust World
50mEthical Machine Learning
27mCrouching Harm, Hidden Bias—Addressing issues of fairness....
51mDigital Discrimination: Cognitive Bias in Machine Learning
38mSocial Implications of Bias in Machine Learning
47mEthical Machine Learning