Roman Lutz
Roman Lutz
We have ongoing work with the scanner feature (`pyrit.cli`) that supports this scenario. A basic version exists and we're working on making it reliable.
#757 did a good chunk of this, but not numpy upgrade. Leaving this open for now.
Anyone can add to this, but we try to focus on things that users request. Group fairness was what we started with, but if you want to add other kinds...
Absolutely. I'm going to put out a PR with an update to markdown files (possibly README, but maybe also others) that shows clearly that we encourage participation and possible projects...
#462 explains this a little bit (mentioning other types of fairness). @riedgar-ms FYI in case you want to elaborate on this more in your PR(?) In general I think we're...
I understand this as a question that turned into the realization that we should add/improve documentation including: - what types of fairness are there (group vs. individual vs. ?) -...
Thanks @StefanieSenger ! I don't have much knowledge of this part of the code. Maybe someone else from @fairlearn/fairlearn-maintainers can take a look?
Thanks for sharing! I may have a different mental model of this than others but in my mind this maps almost 1-to-1 as follows: - “try this right now” —>...
@jai-deepsource is there anything you'd like us to change/update? Happy to make adjustments but it's not clear to me.
Whoever wants to continue this can probably pick up where #506 left off. Might not be too difficult given that there's almost everything there already and it's pre-reviewed, so comments...