notesum.ai
Published at November 4The Intersectionality Problem for Algorithmic Fairness
cs.LG
cs.AI
cs.CY
68T37, 60H30, 68W40
I.6.4; I.5.2
Released Date: November 4, 2024
Authors: Johannes Himmelreich1, Arbie Hsu2, Kristian Lum3, Ellen Veomett2
Aff.: 1Syracuse University; 2University of San Francisco; 3University of Chicago

| Dataset | Subgroup | Subgroup Category | n | |
|---|---|---|---|---|
| adult | (White, Male, 50’s) | [race, sex, age_cat] | 4256 | 0.8020050125313280 |
| bank | <=24 | age_cat | 809 | 0.7766790276060980 |
| compas | (Male, Native American, 25 - 45) | [sex, race, age_cat] | 6 | 0.9444444444444450 |
| compas_violent | (Female, African-American, 25) | [sex, race, age_cat] | 95 | 0.9929824561403510 |
| creditg | (male div/sep, male, <=25) | [personal_status, sex, age_cat] | 2 | 0.6666666666666670 |
| default_credit | (male, 40’s) | [sex, age_cat] | 2771 | 0.8078912546613740 |
| heart_disease | (female, >54) | [sex, age_cat] | 103 | 0.9158576051779940 |
| meps19 | (White, 80’s, female) | [RACE, age_cat, SEX] | 184 | 0.947463768115942 |
| meps20 | (Non-White, 80’s, female) | [RACE, age_cat, SEX] | 146 | 0.9474885844748860 |
| meps21 | (Non-White, 80’s, female) | [RACE, age_cat, SEX] | 142 | 0.9577464788732400 |
| nlsy | (Female, <18, GREEK) | [gender, age_cat, race] | 2 | 0.666666666666667 |
| nursery | great_pret | parents | 4320 | 0.8922839506172840 |
| ricci | W | race | 68 | 0.9901960784313730 |
| student_math | (M, <18) | [sex, age_cat] | 134 | 0.9676616915422890 |
| student_por | (M, >=18) | [sex, age_cat] | 73 | 0.9406392694063930 |
| tae | 1.0 | whether_of_not _the_ta_is_a_native _english_speaker | 29 | 0.8045977011494250 |
| titanic | (female, 60’s) | [sex, age_cat] | 10 | 0.9666666666666670 |
| us_crime | TRUE | blackgt6pct | 970 | 0.9663230240549830 |