#debiasing

[mirrorbreaks]

god that main synth is so addicting

[e-stranger]

we need to debiase databases to make them less sexist

As neural networks tease apart the structure of language, they are finding a hidden gender bias that nobody knew was there.

It is possible to query the vector space of a database to find word embeddings: “Paris : France :: Tokyo : x” will give you the answer x = Japan.

But ask the database “father : doctor :: mother : x” and it will say x = nurse. And the query “man : computer programmer :: woman : x” gives x = homemaker.

But this can be debiased significantly.

One perspective on bias in word embeddings is that it merely reflects bias in society, and therefore one should attempt to debias society rather than word embeddings,” say Bolukbasi and co. “However, by reducing the bias in today’s computer systems (or at least not amplifying the bias), which is increasingly reliant on word embeddings, in a small way debiased word embeddings can hopefully contribute to reducing gender bias in society. https://www.technologyreview.com/s/602025/how-vector-space-mathematics-reveals-the-hidden-sexism-in-language/

This article is from 2016. A more recent article (2019) Lipstick on a Pig:Debiasing Methods Cover up Systematic Gender Biases in Word Embeddings But do not Remove Them by Hila Gonen and Yoav Goldberg from the Department of Computer Science, at the Bar-Ilan University in Tel Aviv suggest that the problem still exists.

The gender bias information is still reflected in the distances between “gender-neutralized” words in the debiased embeddings, and can be recovered from them. We present a series of experiments to support this claim, for two debiasing methods. We conclude that existing bias removaltechniques are insufficient, and should not betrusted for providing gender-neutral modeling. https://arxiv.org/pdf/1903.03862.pdf

[just-evo-now]

Ask meme: send me papers on debiasing.

[altrunews]

Political bias is likely to be a major cause of misguided policies in democracies ( even the main one according to economist Bryan Caplan ).

People are better at judging politically controversial issues in a detached, scientific way today than they were in, say, the 14 th century.

Here I want to focus on another method, however, namely a political bias test which I have constructed with ClearerThinking run by EA-member Spencer Greenberg.

[consultingrationalist-blog]

Intelligence vs. discipline

Francisco Dao wrote for PandoDaily yesterday on grit and discipline:

The entire self improvement industry and what appears to be half of my Facebook newsfeed consist of similar proclamations of “follow your dreams!”

More than just annoying, these messages are flat out wrong and in most cases counterproductive. They teach people to be impulsive and focus solely on the desired outcome while ignoring the path to get there. In study after study, it’s been shown that discipline and impulse control are the primary traits of successful people and the best predictor of future achievement. In contrast, unwarranted, overinflated self esteem is often a sign of future failure. Compared to discipline, even legitimate measures of intelligence have very little bearing as predictors of success.

Indeed, there a few too many intelligent people who choose to indulge in video games rather than apply their gift in life. Speaking from experience, a lot of them became intelligent in the first place through hard work and discipline in another area that they were passionate about. Let's not stray.