4 ms·
Hire minorities to work on your machine learning models. At the very least it ensures you won't 'forget' to train the model with black faces.
by aiaiai 6y ago
Hire minorities to work on your machine learning models. At the very least it ensures you won't 'forget' to train the model with black faces.
- MaximumYComb 6y agoI'm not from the US but there is a link between socioeconomic status and education. I would assume that with the large percentage of minorities, especially black Americans, raised in poverty that there would be a lack of suitable candidates. Just one of the ways that history of racism affects the modern world. Do they have any practical suggestions on where to find a huge number of suitable qualified people? The people in prestigious academic positions often come from not just non-poverty backgrounds but wealthy backgrounds.
- aiaiai 6y agoFacebook is rich enough to invest in black people's education and essentially train data scientists from the ground up. A black scholarship for data science would be a great gesture from them. Plenty of other industries do this for their fields as well.
- lurkmurk 6y agoWell educated people who can work on this problem are not minorities, i.e. they have access to education, higher social status, better starting point. While their color of skin might be different, they don't need to be aware of all the issues their race/sex faces. Example can be difference between treatment of certain minorities in US (where most of the science is developed) and the rest of the world. A scientist at Facebook/Google will develop something for the whole world having US centric view, without access to views from other cultures and their reqiurements for fairness.