segmented.netlify.com points a segmentation model at a set of 100 influential photographs curated by time magazine, and lets you explore what the segmentation mask sees, and what photographs are behind it
This is absolutely brilliant work! ❤️
imho it's up there with some of the best work on CS education👇
Epistemological Pluralism
@STurkle & Papert 1990
Stuck in The Shallow End
@joannagoode13 & pals 2010
You Can't Sit With Us
@rajiinio@morganklauss@amironesei 2021
💻❤️😎
🗣️Finally, it's out! It's here!
In this paper, me, @morganklauss & @amironesei present work on a topic people remain hesitant to discuss - the **power dynamic between disciplines** & how that shapes poor AI ethics education and practice.
Check it out: dl.acm.org/doi/10.1145/34…
"a big part of the problem has to do with labeling... it's this cultural aspect that is giving this problem a very tricky dimension"
"i've learned... it's very important to consider these cultural aspects of language before considering that the model is performing well"
Toxic Language Detection is a problem that is *much* harder than I thought it would be and also for a reason that I did not anticipate. To those interested; here's a video of me building such a system and spotting a flaw on @Rasa_HQ algorithm whiteboard.
youtube.com/watch?v=Czto6G…
#emnlp "Stop assuming that the future will look like the past, because it completely discounts everything that marginal communities are currently doing and are currently experiencing. It essentially negates the potential for human evolution."
1/n @HousingTheCity
"it's a crisis in public understanding in science... ML basically impacts everyone everywhere... yet so many times folks in the ML community are not communicating or collaborating"
@hannawallach@dxmartinjr "it's a crisis in science understanding the public"
#PAIRSymposium
