3 ms·

The researcher said she identified racial bias. The factors you describe are merely reasons why such bias would exist. What’s your basis for saying this researc

The researcher said she identified racial bias. The factors you describe are merely reasons why such bias would exist. What’s your basis for saying this researcher is race baiting, hyper partisan, or showing a lack of understanding of anything?

FakeComments 8y ago
A racial bias is not indicative of racism — and is not inherently problematic. The entire tone of the article, and indeed her own comments, is that this racial bias is somehow problematic or must be gotten rid of. > What’s struck a chord about Boulamwini’s work is her method of testing the systems created by well-known companies. She applies such systems to a skin-tone scale used by dermatologists, then names and shames those that show racial and gender bias. Buolamwini, who’s also founded a coalition of scholars, activists and others called the Algorithmic Justice League, has blended her scholarly investigations with activism. The question of differential identification between white and black may merely be a question of physics (ie, any recognition system has an easier time of white people because of how optics work), and the relevant question for implementing facial recognition is if it works well enough at identifying black people to be usable. To name and shame people because their product responds to genuine physical differences in groups of people is absurd. Comparing white people and black people without acknowledging the racial disparity in optical response is race baiting; comparing men and women without acknowledging the broad scale difference in bone structure is gender baiting. Why? Because there is no problem with a racial or gender bias if it extends from a physiological difference, and not willful discrimination. Her research is the equivalent of concluding the grocery store is problematically heightist because tall people have an easier time reaching the top shelf — unseriousness unworthy of the topic, and outrage bait.
- tzar 8y ago
  Isn't the problem, then, that the software doesn't work well enough with dark skin?
  - FakeComments 8y ago
    I’ve seen no evidence of it. What I’ve seen is differential comparisons, eg, comparing the rate of white detection to black detection or the difference in certainty scores on each — but I’d really appreciate it if people could show me the actual certainty numbers on black faces so I can see if it’s failing to recognize, misrecognizing, or just less sure then white faces.
    - got2surf 8y ago
      The evidence is literally in the original article: "Darker-skinned women were the most misclassified group, with error rates of up to 34.7%. By contrast, the maximum error rate for lighter-skinned males was less than 1%" Regardless of whether the higher error rate is a combination of race, gender, or both, it's still a huge issue. Granted, that study was from a year ago, and other companies have since improved their facial recognition systems. But an overall precision/accuracy/f1 score doesn't mean much when accuracy varies that much by group. Sure, you can market it as "accurate on white males", but you can't market it as "accurate"
- mikeash 8y ago
  The article contains no occurrences of the term “racism” or “racist.” Why is there necessarily no problem with a bias if it comes from a physiological difference? If a certain group of people gets falsely convicted of crimes at a disproportionate rate because some computer program misidentifies them more frequently than it does other groups, that is a huge problem that needs to be fixed, even if the misidentification is due to some inherent difference between the group.
  - FakeComments 8y ago
    You are correct — however, it mentions her politicking on the basis of it being so: > Buolamwini, who’s also founded a coalition of scholars, activists and others called the Algorithmic Justice League, has blended her scholarly investigations with activism. To respond to your main point: > If a certain group of people gets falsely convicted of crimes at a disproportionate rate because some computer program misidentifies them more frequently than it does other groups, that is a huge problem that needs to be fixed No, it doesn’t. Not if the reason it’s happening is some physical characteristic just makes it harder to be sure in those cases. The only relevant question is if the rate in each category is acceptable in terms of false positives, and truly, if in each case we reach a level of certainty that convicts the individual beyond a reasonable doubt. The only reason that you care so much is that you’re a racist, who believes we need to balance the treatment of racially essential tribes — rather than that we need to ensure our technology works at an appropriate level on an individual basis, where varying physical characteristics mean it works more or less well for different people, for which the only relevant criterion for its use in a case is its certainty for a particular face (and not its relative certainty for classes of faces). I notice you people only seem to be outraged on the topic of small differences in skin color — but that familial DNA tracing works better on men than women is just accepted and taken as normal. Can you explain why you think differential facial recognition scores actually is meaningful? I mean, you asserted it was a problem — what particularly is problematic about it? Edit: For below — no, just this particular case isn’t racism, it’s a genuine difference between people and someone trying to racebait about it. Not every racial difference is racism, and a lot of claims of racism are just made up nonsense. https://youtu.be/wR2lK36Tc4M https://youtu.be/wR2lK36Tc4M
golemiprague 8y ago
Bias is something intentional or at least stems from some wrong thinking, even sub conscious. The point here is that it is not bias but rather technical reasons. Not every difference is bias. There is a common trend now to look at every difference in outcome as bias or discrimination and it is very political motivated, works only to one direction and in general has no evidence except from the different outcome itself which could be explained by million reasons. Difference is not bias.

FakeComments 8y ago

A racial bias is not indicative of racism — and is not inherently problematic. The entire tone of the article, and indeed her own comments, is that this racial bias is somehow problematic or must be gotten rid of. > What’s struck a chord about Boulamwini’s work is her method of testing the systems created by well-known companies. She applies such systems to a skin-tone scale used by dermatologists, then names and shames those that show racial and gender bias. Buolamwini, who’s also founded a coalition of scholars, activists and others called the Algorithmic Justice League, has blended her scholarly investigations with activism. The question of differential identification between white and black may merely be a question of physics (ie, any recognition system has an easier time of white people because of how optics work), and the relevant question for implementing facial recognition is if it works well enough at identifying black people to be usable. To name and shame people because their product responds to genuine physical differences in groups of people is absurd. Comparing white people and black people without acknowledging the racial disparity in optical response is race baiting; comparing men and women without acknowledging the broad scale difference in bone structure is gender baiting. Why? Because there is no problem with a racial or gender bias if it extends from a physiological difference, and not willful discrimination. Her research is the equivalent of concluding the grocery store is problematically heightist because tall people have an easier time reaching the top shelf — unseriousness unworthy of the topic, and outrage bait.

tzar 8y ago
Isn't the problem, then, that the software doesn't work well enough with dark skin?
- FakeComments 8y ago
  I’ve seen no evidence of it. What I’ve seen is differential comparisons, eg, comparing the rate of white detection to black detection or the difference in certainty scores on each — but I’d really appreciate it if people could show me the actual certainty numbers on black faces so I can see if it’s failing to recognize, misrecognizing, or just less sure then white faces.
  - got2surf 8y ago
    The evidence is literally in the original article: "Darker-skinned women were the most misclassified group, with error rates of up to 34.7%. By contrast, the maximum error rate for lighter-skinned males was less than 1%" Regardless of whether the higher error rate is a combination of race, gender, or both, it's still a huge issue. Granted, that study was from a year ago, and other companies have since improved their facial recognition systems. But an overall precision/accuracy/f1 score doesn't mean much when accuracy varies that much by group. Sure, you can market it as "accurate on white males", but you can't market it as "accurate"
mikeash 8y ago
The article contains no occurrences of the term “racism” or “racist.” Why is there necessarily no problem with a bias if it comes from a physiological difference? If a certain group of people gets falsely convicted of crimes at a disproportionate rate because some computer program misidentifies them more frequently than it does other groups, that is a huge problem that needs to be fixed, even if the misidentification is due to some inherent difference between the group.
- FakeComments 8y ago
  You are correct — however, it mentions her politicking on the basis of it being so: > Buolamwini, who’s also founded a coalition of scholars, activists and others called the Algorithmic Justice League, has blended her scholarly investigations with activism. To respond to your main point: > If a certain group of people gets falsely convicted of crimes at a disproportionate rate because some computer program misidentifies them more frequently than it does other groups, that is a huge problem that needs to be fixed No, it doesn’t. Not if the reason it’s happening is some physical characteristic just makes it harder to be sure in those cases. The only relevant question is if the rate in each category is acceptable in terms of false positives, and truly, if in each case we reach a level of certainty that convicts the individual beyond a reasonable doubt. The only reason that you care so much is that you’re a racist, who believes we need to balance the treatment of racially essential tribes — rather than that we need to ensure our technology works at an appropriate level on an individual basis, where varying physical characteristics mean it works more or less well for different people, for which the only relevant criterion for its use in a case is its certainty for a particular face (and not its relative certainty for classes of faces). I notice you people only seem to be outraged on the topic of small differences in skin color — but that familial DNA tracing works better on men than women is just accepted and taken as normal. Can you explain why you think differential facial recognition scores actually is meaningful? I mean, you asserted it was a problem — what particularly is problematic about it? Edit: For below — no, just this particular case isn’t racism, it’s a genuine difference between people and someone trying to racebait about it. Not every racial difference is racism, and a lot of claims of racism are just made up nonsense. https://youtu.be/wR2lK36Tc4M https://youtu.be/wR2lK36Tc4M

golemiprague 8y ago

Bias is something intentional or at least stems from some wrong thinking, even sub conscious. The point here is that it is not bias but rather technical reasons. Not every difference is bias. There is a common trend now to look at every difference in outcome as bias or discrimination and it is very political motivated, works only to one direction and in general has no evidence except from the different outcome itself which could be explained by million reasons. Difference is not bias.