I think you're a bit off the mark with this one. What someone means by bias is probably obvious from context. Moreover, the isolated view of bias in all forms as group-level discrimination is probably more characteristic of critical theory than sociology.
Four different definitions of bias, not to be confused: Statistics: Difference btwn correct answer & average one. ML: Prior knowledge without which learning is impossible. Psychology: Flip side of cognitive heuristics. Sociology: Racism, sexism, etc.
-
-
-
It's usually obvious, but popular reporting on algorithms is a very important exception:https://jacobitemag.com/2017/08/29/a-i-bias-doesnt-mean-what-journalists-want-you-to-think-it-means/ …
Keskustelun loppu
Uusi keskustelu -
-
-
The objective is not to minimize bias, but to minimize error. There is a lesson to be learned here.
Kiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
-
-
https://www.collinsdictionary.com/dictionary/english/bias … 1. NOUN Bias is a tendency to prefer one person or thing to another, and to favour that person or thing 2. NOUN Bias is a concern with or interest in one thing more than others 3. VERB To bias someone means to influence them in favour of a specific choice
-
Correct.
Keskustelun loppu
Uusi keskustelu -
-
-
I have read about it, but... Which are the real life consequences of introducing bias in Machine Learning?
-
ML = bias. Continue this sequence: "1 2 3 4". You pick 5, choosing f(x)=x, instead of picking 22, choosing a very complicated polynomial for f(x). So you're "biased" towards low-degree polynomials, but there's no apriori reason 5 is a better answer than 22.
- Näytä vastaukset
Uusi keskustelu -
-
-
ML and human psychological biases are not as different as (even thoughtful) people think.
Kiitos. Käytämme tätä aikajanasi parantamiseen. KumoaKumoa
-
-
-
How about parametrising the outcome of human beings live’s by parameters they hold no power over such as their race, sexual preference or gender? ML models mirror the goods and the bads of human decisions and some mitigation of the bads is not bad per se
-
Example: just because women are the standard in kindergarten teaching positions doesn’t mean that a man cannot become a brilliant kindergarten teacher. However if we make an ML model to choose a kindergarten teacher, as the man is the outlier, it will end up excluding him
Keskustelun loppu
Uusi keskustelu -
Lataaminen näyttää kestävän hetken.
Twitter saattaa olla ruuhkautunut tai ongelma on muuten hetkellinen. Yritä uudelleen tai käy Twitterin tilasivulla saadaksesi lisätietoja.