Algorithm Club
@algorithm_club
Twitter book club on the theme of critical algorithm studies. Moderated by @hyper_linda and @tadiseshan
You might like
The first thing on the reading list for @algorithm_club is the Code of Conduct: github.com/drtortoise/cri…
#UX people - apply to go to Barcelona to work on UX + algorithmic bias this summer. @carolinesinders @taraadiseshan @algorithm_club
Open Call // Artistic residence on "Interface Bias" @Hangar_org // + Info bit.ly/2hAjeh7 #criticalInterface #algorithmicBias
Love this list of critical thoughts on tech by women! Hopefully we can draw some future readings from here.
That list of women tech critics just got a lot longer thanks to a pull request from Tim Chevalier, thank you! zararah.net/blog/2016/03/1…
of course we have the most influence over algs we help create, but the ProPub model is very powerful #algclub
Corollary: have ways to update algo, in software, research, & public eye. Not everyone reads academic journals. #algclub
Always analyze algos' results. Have goals. An algorithm isn't set in stone when it's first released. #algclub
we have to pull back the curtain and de-abstract the data, show it's the result of people typing #algclub
Time for us to wrap up! thank you all for joining! we definitely forgot it was a public holiday. :) #algclub
@hyper_linda @meetar @algorithm_club public data is one areas where its immediately obvious that data is redacted/missing
every single one of these questions is actually a riddle #algclub
Q5: What can we, as folks who work with data, do to help investigate and hold algorithmic systems accountable? #algclub
Q4: the idea that data is always, always, *by definition* incomplete should be the default #algclub
I make maps, so my first Q of any dataset is always: what's been left out? But this assumption is not the default #algclub
Libraries are way ahead of many on this front – in 03 the ALA advocated not saving any unnecessary info #algclub
A4: in the dataset, collecting more info (demographics) gives more context. But opens up data to later abuse. #algclub
A3: I roflmao that Congress & TSA thought everyone would opt into Precheck (which requires finger printing & more) #algclub
Q4: How can we provide context around biases in collection and data usage in the dataset itself? #algclub
Also it's important to recognize that there are places where data collection is mandatory, like immigration. #algclub
so algorithmic bias feeds back into user bias, this is why it's called systemic :/ #algclub
United States Trends
- 1. Grammy 350K posts
- 2. Dizzy 9,987 posts
- 3. Clipse 20.9K posts
- 4. #FliffCashFriday N/A
- 5. Kendrick 62.6K posts
- 6. #NXXT 1,050 posts
- 7. #GOPHealthCareShutdown 7,311 posts
- 8. James Watson 4,069 posts
- 9. Orban 39.4K posts
- 10. #FursuitFriday 12.2K posts
- 11. addison rae 24.3K posts
- 12. olivia dean 15.7K posts
- 13. Leon Thomas 20.3K posts
- 14. Katseye 119K posts
- 15. AOTY 21.9K posts
- 16. Darryl Strawberry N/A
- 17. ravyn lenae 4,686 posts
- 18. Thune 72.4K posts
- 19. Alfredo 2 1,056 posts
- 20. Carmen 48.7K posts
Something went wrong.
Something went wrong.