Social Implications of Bias in Machine Learning
summary
Shows concrete failures and fixes: a pothole app finds only wealthy neighborhoods until buses are instrumented; image search mislabels cats and hides why until LIME highlights influential pixels; Google Translate flips gender until supplemental data is added; facial recognition and COMPAS show much higher error rates for people of color. Recommends tools and practices: Data Ethics Canvas, Facets, the What‑If tool, auditor questions and intentional data collection to avoid feedback loops and to document defaults, thresholds and provenance.
AI-written summary · about
chapters (4)
- 0:00Start
- 1:18Statistical bias Average of measurements vs reality
- 2:33Algorithms and statistical models that computer systems use to effectively perform a specific task, without using explicit instructions, relying on patterns and inference instead
- 11:27Minimising negative impact?
related talks
38mSocial Implications of Bias in Machine Learning
47mEthical Machine Learning
37mKeynote: Ethical Machine Learning: Creating Fair Models in an Unjust World
27mCrouching Harm, Hidden Bias—Addressing issues of fairness....
51mDigital Discrimination: Cognitive Bias in Machine Learning
47mEthics & AI: Identifying & preventing bias in predictive models
54mEthical Machine Learning
50mEthical Machine Learning