• 0 Posts
  • 10 Comments
Joined 9 months ago
cake
Cake day: September 27th, 2023

help-circle




  • Mirodir@discuss.tchncs.detoLemmy Shitpost@lemmy.worldAutomation
    link
    fedilink
    arrow-up
    14
    arrow-down
    1
    ·
    10 days ago

    So is the example with the dogs/wolves and the example in the OP.

    As to how hard to resolve, the dog/wolves one might be quite difficult, but for the example in the OP, it wouldn’t be hard to feed in all images (during training) with randomly chosen backgrounds to remove the model’s ability to draw any conclusions based on background.

    However this would probably unearth the next issue. The one where the human graders, who were probably used to create the original training dataset, have their own biases based on race, gender, appearance, etc. This doesn’t even necessarily mean that they were racist/sexist/etc, just that they struggle to detect certain emotions in certain groups of people. The model would then replicate those issues.





  • but we did damage a 5000-year-old monument

    As far as I could find out, they used orange cornflour that will just wash off the next time it rains. The most amount of damage anyone could seriously bring up was that it could harm/displace the lichen on the henge.

    That’s not to say that I specifically condone the action, but it’s a lot less bad than this article makes it sound. It’s the same with the soup attack on one of van Gogh’s painting, which had protective glass on it. So far all the JSO actions targeting cultural/historical things (at least the ones that made it to the big news) have been done in a way that makes them sound awful at first hearing, but intentionally did not actually damage the targeted cultural/historical thing.

    I think the biases of the journalist/news outlet/etc. are somewhat exposed by which parts they focus on and which they downplay or omit entirely.


  • Also if we give it the benefit of the doubt (and it really is a stretch to make this work lol): I could make the argument that this person meant to write: “The movie has such a terrible premise, yet it was successful enough to have two sequels. Learning how it got that success despite the material’s premise taught me these 5 things about product management:” and just worded it terribly.