Update README.md

firmai · Mar 31, 2020 · 3460a17 · 3460a17
1 parent e4eb1cf
commit 3460a17
Showing 1 changed file with 18 additions and 18 deletions.
diff --git a/README.md b/README.md
@@ -67,24 +67,24 @@ What Questions Do We Attempt to Answer?
 ------------
 
 1. Can the model predict the outcome using just protected values? (Protected Value Prediction)
-1. Is the model monotonic and are variables randomly selected? (Model Constraints, LV1 & LV2 Monotonicity)
-1. Is the model explainable? (Model Selection, Feature Interactions)
-1. Can you explain the predictions globally and locally? (SHAP)
-1. Does the model perform well? (Metrics)
-1. What indivuals have received the most and least accurate predictions? (Residual Deviation)
-1. Can you point to the feature responsible for large individual residuals? (Residual Explanations)
-1. What feature values could potentially be outliers due to their misprediction? (Residual Explanations)
-1. Do some models perform better at predicting the outcomes for a certain type of individual? (Benchmark Competition)
-1. Can the model outcome be changed by artificially perturbing certain values of interest? (Adverserial Attack)
-1. Do certain groups suffer relative to others as measured through group statistics? (Parity Indicators, Fair Lending Measures)
-1. Can various data and prediction processing techniques improve these group statistics? (Model Agnostic Processing)
-1. What features are driving the structural differences between groups controlling for demographic factors? (Feature Decomposition)
-1. What inviduals have received the most unfair prediction or treatment by the model? (Individual Disparity)
-1. Why did the model decide to predict a specific outcome for a particular individual or sub-group of individuals? (Reasoning Codes)
-1. What indviduals are most similar to those receiving unfair treatment and were these indivduals treated similary? (Prototypical)
-1. What individual is the closest related instance to a sample individual but has a different predicted outcome? (Counterfactual)
-1. What is the minimal feature pertubation necessary to switch an individual's prediction to another category? (Contrastive)
-1. What is the maximum perturbation possible while the model prediction remains the same? (Contrastive)
+2. Is the model monotonic and are variables randomly selected? (Model Constraints, LV1 & LV2 Monotonicity)
+3. Is the model explainable? (Model Selection, Feature Interactions)
+4. Can you explain the predictions globally and locally? (SHAP)
+5. Does the model perform well? (Metrics)
+6. What individuals have received the most and least accurate predictions? (Residual Deviation)
+7. Can you point to the feature responsible for large individual residuals? (Residual Explanations)
+8. What feature values could potentially be outliers due to their misprediction? (Residual Explanations)
+9. Do some models perform better at predicting the outcomes for a certain type of individual? (Benchmark Competition)
+10. Can the model outcome be changed by artificially perturbing certain values of interest? (Adversarial Attack)
+11. Do certain groups suffer relative to others as measured through group statistics? (Parity Indicators, Fair Lending Measures)
+12. Can various data and prediction processing techniques improve these group statistics? (Model Agnostic Processing)
+13. What features are driving the structural differences between groups controlling for demographic factors? (Feature Decomposition)
+14. What individuals have received the most unfair prediction or treatment by the model? (Individual Disparity)
+15. Why did the model decide to predict a specific outcome for a  particular individual or sub-group of individuals? (Reasoning Codes)
+16. What individuals are most similar to those receiving unfair treatment and were these individuals treated similar? (Prototypical)
+17. What individual is the closest related instance to a sample individual but has a different predicted outcome? (Counterfactual)
+18. What is the minimal feature perturbation necessary to switch an individual's prediction to another category? (Contrastive)
+19. What is the maximum perturbation possible while the model prediction remains the same? (Contrastive)