Week 31: Datasets for fairness

This week, Diana and I looked for some datasets on which to evaluate our fairness criteria. This is no simple task. For one, the datasets need to be large enough to be split into training, calibration, and testing sets. Second, there needs to be some protected group attribute like sex, race, or age. Third, there needs to be some outcome ranking as well as some truth value. The truth is often the hardest attribute to find.

In this post, I introduce some of the datsets we decided to use. I try to identify each of the necessary features that we need to properly test our fairness criteria.

COMPAS

The COMPAS dataset is an obvious choice because it is well ingrained in the literature about fairness. It has 18,000 observations (n=18000) and 1000 attributes (k=1000). Protected groups can be sex (81/19) or race (51/49). The outcome attribute can be the recidivism score, the score used in court that represents the risk that the accused will commit another crime. The truth attribute is a little difficult; the number of months until they commit another crime? The severity of their future crime, should they commit one?

German Credit

This is a dataset used in the FA*IR paper from CIKM last year. n=1000 and k=100, and groups can be sex (69/31) or age (55/45). The outcome attribute is the calculated schufa score, and the truth attribute can be the amount of time before they default on a loan in the future? The amount on which they default? Unfortunately, this future data is not readily available to us.

SAT Score

This is another datset from FA*IR. n=1.6m and k=1500, and sex is a protected attribute (53/47). The outcome attribute is SAT score and truth attribute could be college GPA? Starting salary at their first job?

Summary

We identified 11 such datasets, each of which we are unsure what the truth attribute can be. For purposes of testing, we can just choose one attribute to represent the truth. However, if we want to talk about these datasets in a paper, we want to make sure that the story is compelling and that it helps show the utility of our fairness correction methods.

In the following week, Diana and I will be working on cleaning up these datasets and fixing some of the Python scripts to better streamline our testing process. Until then!

Comments

Post a Comment

Popular posts from this blog

Week 32-33: Wrapping up

Hi all. This is my last blog post for this project. The last few weeks, we spent little time working on the project and instead focused on finishing up our classes. To be honest, we lost some motivation after finishing the paper, and we've been reluctant to start up new tasks so close to the end of the year. In this post, I'd like to talk about some of the things I've learned in this project other than what we found in our research. This post is really more a service to myself; to help me reflect on this past year and what I've learned. When I started this project, I didn't understand how research in computing worked. My exposure to research had been more biology/clinical, in which research involved setting a hypothesis about some protein, designing an experiment with positive and negative controls, and then running the experiment repeatedly until confident in the results. In computing research, my experience has been of two varieties: reading papers and proce...
Read more

Week 30: Returning to Fairness

With the LineUp paper submitted, it's now time to return to the fairness problem we were working on before. As a bit of a refresher for you and for me, I'll review what we've done with fairness so far and what remains to be done. Fairness in Ranking Up until now, the literature regarding fairness has focused on fairness in classification. As an illustrative example, consider fairness in hiring practices with repect to applicant sex. A job opening might receive 100 applicants evenly split into 50 males and 50 females. If males and females are equally capable of doing the job, then you might be concerned to see the hiring company mostly interviewing male applicants. To evaluate fairness, you would compare each applicant's capability (capable of performing duties, not capable) to their outcome (interviewed, not interviewed). In a completely fair scenario, the same proportion of capable applicants from each sex group would receive an interview. Small deviances from this...
Read more