I find this topic incredibly interesting and of paramount importance.
Here a list of links to papers/repos/tools I am currently going through to understand how to check for bias in your datasets/algorithms AND address this bias:
- https://arxiv.org/pdf/1810.01943.pdf
- https://arxiv.org/pdf/1901.10002.pdf
- https://arxiv.org/ftp/arxiv/papers/1907/1907.09013.pdf
- http://www.datasciencepublicpolicy.org/projects/aequitas/
- https://dssg.github.io/aequitas/metrics.html#Preliminary-Concepts
- https://github.com/dssg/aequitas
- https://arxiv.org/pdf/1901.04966v1.pdf
- https://github.com/ibm/aif360
- https://aif360.mybluemix.net/
- https://www.youtube.com/watch?v=X1NsrcaRQTE
- https://doteveryone.org.uk/wp-content/uploads/2019/06/AIES-19_paper_223.pdf
AIF360 (ironically from IBM…) is the most complete tool I have explored so far.