It's pretty impossible for ML not to have bias
The only ML algorithms with race and gender biases are those concocted by "ethical AI" researchers.
-
-
Kiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
-
-
Tämä twiitti ei ole saatavilla.
-
No it's not ok in fact.
Keskustelun loppu
-
-
-
can machines really learn without bias? to learn is to generalize to generalize is to make a judgement call
Kiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
-
-
Apple's safe browsing mode blocks sites with the word "Asian" in them. This is what happens when you train on white male coders' browsing histories.
-
That being said, this is more of an issue with training datasets than the algorithms themselves. SGD doesn't know what race is.
Keskustelun loppu
Uusi keskustelu -
-
-
Kiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
-
-
Deciding what to correct for does introduce bias. But when narrow datasets amplify bias, that's worse. Of course, deciding which is worse introduces bias. Validating your original point.
Kiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
-
-
Deep learning models don't have willful ignorance and the bias of wokeness. So they discover true patterns in the real world data such as that Africans have genetic traits that predispose them to lower IQs, are built for violence, and are behaviorally predisposed to violence.
-
And that Asians have the opposite genetic traits which are expressed in their physiology and behavior. The great thing about nn models is that we can rediscover the truth again without current human biases.
- Näytä vastaukset
Uusi keskustelu -
Lataaminen näyttää kestävän hetken.
Twitter saattaa olla ruuhkautunut tai ongelma on muuten hetkellinen. Yritä uudelleen tai käy Twitterin tilasivulla saadaksesi lisätietoja.