It’s Not Fair! Detecting Algorithmic Bias with Open Source Tools
summary
Practical methods for spotting and assessing algorithmic bias using open-source tools are presented, focusing on measuring when predictive models reproduce or amplify unfairness. The session covers how to combine available open-source libraries, an investigative mindset, and a structured testing strategy to evaluate algorithmic impacts in real-world systems such as recommender systems and predictive risk models.
AI-written summary · about
chapters (9)
- 0:00Start
- 10:11Impact of Transparency
- 16:04COMPAS Model Interpretability
- 20:06Model Agnostic Methods
- 21:54Comparison of Transparency Tools
- 28:00LIME Results: Diabetes Dataset
- 34:32Hidden Bias: Renal Failure Dataset
- 35:43Aequitas Fairness Toolkit
- 38:08Aequitas: Formatting the Data Set
related talks
35mOpen Source Fairness
51mDigital Discrimination: Cognitive Bias in Machine Learning
37mKeynote: Ethical Machine Learning: Creating Fair Models in an Unjust World
27mCrouching Harm, Hidden Bias—Addressing issues of fairness....
40mQuestions and Practices to Make Algorithmic Decision Making More Fair
23mRemoving Unfair Bias in the Machine Learning Pipeline
37mDigital Discrimination: Cognitive Bias in Machine Learning
40mEthical AI: Addressing Bias & Algorithmic Fairness