#bias

You can’t characterize human nature if studies overlook 85 percent ...
▻https://diasp.eu/p/8487211

You can’t characterize human nature if studies overlook 85 percent of people on Earth | #bias #psychology #research #WEIRD

Bias in #ai: a letter to my daughter
▻https://hackernoon.com/my-dearest-eliisabet-de724a86d105?source=rss----3a8144eabfe3---4

https://cdn-images-1.medium.com/max/1024/1*80RJcYHioUimZkIlN_VeeQ.jpeg

My dearest Eliisabet,Only 5 years old, but you will grow up to become a strong, beautiful & independent woman, of that I’m sure.You are a daughter of a man that wasn’t always an example to others. It took some time for me to grow up & take responsibility for my actions. Being your father has accelerated this process for sure.My generation seems to be struggling with growing up as well. Especially in the relatively young tech community, and more specific the artificial intelligence community, people are not always taking their responsibilities serious.Equality is hard work for my non-white & non-male colleagues in both academia and the industry. The current system is biased in many subtle but also in some very obvious ways.Although the tech community has an huge impact on (...)

#feminism #bias-in-ai #ai-bias #women-in-tech

Tackle #bias and Other Problems/Solutions in Machine Learning Models
▻https://hackernoon.com/tackle-bias-and-other-problems-solutions-in-machine-learning-models-f427

Predictive Analytics models rely heavily on Regression, Classification and Clustering methods. When analysing the effectiveness of a predictive model, the closer the predictions are to the actual data, the better it is. This article hopes to be a one-stop reference to the major problems and their most popular/effective solutions, without diving into details for execution.A Linear Regression PlotA clustering algorithm plotPrimarily, data selection and pruning happens during the Data Preparation phase, where you take care to get rid of bad data in the first place. Then again, there are issues with the data, and their relevance to the ML model’s objectives during training, troubles with usage of #algorithms, and errors in the data that occur throughout. Effectively, the model is tested for (...)

#machine-learning #machine-learning-models #predictive-analytics

Belief in FakeNews is Associated with Delusionality, Dogmatism, Rel...
▻https://diasp.eu/p/8000419

Belief in FakeNews is Associated with Delusionality, Dogmatism, Religious Fundamentalism, and Reduced Analytic Thinking | #delusionality #bias #analytic #analyticthinking #belief #fundamentalism #fakenews #thinking #cognition #dogmatism

Don’t Want to Fall for Fake News? Don’t Be Lazy (▻https://www.wired....
▻https://diasp.eu/p/8000416

Don’t Want to Fall for Fake News? Don’t Be Lazy | #bias #analyticthinking #reflexivity #lazy #CRT #cognition #fakenews

Extending the seductive allure of neuroscience explanations effect ...
▻https://diasp.eu/p/7759430

Extending the seductive allure of neuroscience explanations effect to popular articles about educational topics - Im - 2017 - British Journal of Educational Psychology - Wiley Online Library | #psychology #persuasion #influence #bias #neuroscience #education #explanations #seduction

The Akrasia Effect: Why We Don’t Follow Through on Things (https://...
▻https://diasp.eu/p/7545431

The Akrasia Effect: Why We Don’t Follow Through on Things | #akrasia #desire #selfcontrol #time #bias #gratification #intentions #procrastination #friction #strategy #cognition #addiction #inconsistancy #engagement

Quantifying Biases in Online Information Exposure | Center for Complex Networks and Systems Research, Indiana University
▻https://arxiv.org/abs/1807.06958
▻https://arxiv.org/pdf/1807.06958.pdf

Our consumption of online #information is mediated by filtering, ranking, and recommendation algorithms that introduce unintentional biases as they attempt to deliver relevant and engaging content. It has been suggested that our reliance on online technologies such as search engines and social media may limit exposure to diverse points of view and make us vulnerable to manipulation by disinformation. In this paper, we mine a massive dataset of Web traffic to quantify two kinds of bias: (i) homogeneity bias, which is the tendency to consume content from a narrow set of information sources, and (ii) popularity bias, which is the selective exposure to content from top sites. Our analysis reveals different bias levels across several widely used Web platforms. Search exposes users to a diverse set of sources, while social media traffic tends to exhibit high popularity and homogeneity #bias. When we focus our analysis on traffic to news sites, we find higher levels of popularity bias, with smaller differences across applications. Overall, our results quantify the extent to which our choices of online systems confine us inside “social bubbles.”

#personnalisation #médias_sociaux #algorithme via @pomeranian99

Flow chart of cognitive biases (▻https://news.ycombinator.com/item?i...
▻https://diasp.eu/p/7043595

Flow chart of cognitive biases | #bestof #cognitive #biases #cognition

The Trouble with Bias - Kate Crawford - NIPS 2017 (50mn)
▻https://www.youtube.com/watch?v=6Uao14eIyGc&t=3m46s

Great talk about #bias, #fairness and #responsibility in #machineLearning

(and thanks to whoever put those NIPS talks somewhere else than Facebook)

U.S. Black Mothers Die In Childbirth At Three Times The Rate Of White Mothers : NPR
▻https://www.npr.org/2017/12/07/568948782/black-mothers-keep-dying-after-giving-birth-shalon-irvings-story-explains-why

Black women are more likely to be uninsured outside of pregnancy, when Medicaid kicks in, and thus more likely to start prenatal care later and to lose coverage in the postpartum period. They are more likely to have chronic conditions such as obesity, diabetes and hypertension that make having a baby more dangerous. The hospitals where they give birth are often the products of historical #segregation, lower in quality than those where white mothers deliver, with significantly higher rates of life-threatening complications.

Those problems are amplified by unconscious #biases that are embedded in the medical system, affecting quality of care in stark and subtle ways. In the more than 200 stories of #African-American mothers that ProPublica and NPR have collected over the past year, the feeling of being devalued and disrespected by medical providers was a constant theme.

There was the new mother in Nebraska with a history of hypertension who couldn’t get her doctors to believe she was having a heart attack until she had another one. The young Florida mother-to-be whose breathing problems were blamed on obesity when in fact her lungs were filling with fluid and her heart was failing. The Arizona mother whose anesthesiologist assumed she smoked marijuana because of the way she did her hair. The Chicago-area businesswoman with a high-risk pregnancy who was so upset at her doctor’s attitude that she changed OB/GYNs in her seventh month, only to suffer a fatal postpartum stroke.
Over and over, black women told of medical providers who equated being African-American with being poor, uneducated, noncompliant and unworthy. “Sometimes you just know in your bones when someone feels contempt for you based on your #race,” said one Brooklyn, N.Y., woman who took to bringing her white husband or in-laws to every prenatal visit. Hakima Payne, a mother of nine in Kansas City, Mo., who used to be a labor and delivery nurse and still attends births as a midwife-doula, has seen this cultural divide as both patient and caregiver. “The nursing culture is white, middle-class and female, so is largely built around that identity. Anything that doesn’t fit that #identity is suspect,” she said. Payne, who lectures on unconscious bias for professional organizations, recalled “the conversations that took place behind the nurse’s station that just made assumptions; a lot of victim-blaming — ’If those people would only do blah, blah, blah, things would be different.’”
...
But it’s the discrimination that black women experience in the rest of their lives — the double whammy of race and gender — that may ultimately be the most significant factor in poor maternal outcomes.

“It’s chronic stress that just happens all the time — there is never a period where there’s rest from it. It’s everywhere; it’s in the air; it’s just affecting everything,” said Fleda Mask Jackson, an Atlanta researcher who focuses on birth outcomes for middle-class black women.

It’s a type of stress for which education and class provide no protection. “When you interview these doctors and lawyers and business executives, when you interview African-American college graduates, it’s not like their lives have been a walk in the park,” said Michael Lu, a longtime disparities researcher and former head of the Maternal and Child Health Bureau of the Health Resources and Services Administration, the main federal agency funding programs for mothers and infants. “It’s the experience of having to work harder than anybody else just to get equal pay and equal respect. It’s being followed around when you’re shopping at a nice store, or being stopped by the police when you’re driving in a nice neighborhood.”

#racisme #États_Unis #maternité

How white engineers built racist code – and why it’s dangerous for black people

As facial recognition tools plays a bigger role in fighting crime, inbuilt racial biases raise troubling questions about the systems that create them
▻https://www.theguardian.com/technology/2017/dec/04/racist-facial-recognition-white-coders-black-people-police
@etraces @copycult @constant via Marie Lechner.
#algorythmes #biases

How we transferred our biases into our machines and what we can do about it
▻https://youarenotsosmart.com/2017/11/20/yanss-115-how-we-transferred-our-biases-into-our-machines-and-what

three experts on artificial intelligence help us understand how we accidentally transferred our prejudices and biases into our infant artificial intelligences. We will also explore who gets to say what is right and what is wrong as we try to fix all this. And you’ll hear examples of how some of our early machine minds, through prediction, are creating the future they predict by influencing the systems they monitor — because our actions folds their results back into their next prediction.

A great episode (and surprisingly optimistic episode given the subject)

#ai #ethics #bias #podcast

The Words Men and Women Use When They Write About Love

https://framapic.org/w5MQkSHY3Wq3/0bJyUYscWW2h.png

We ignore what doesn’t fit with our biases – even if it costs us (h...
▻http://02mydafsoup-01.soup.io/post/631540059/We-ignore-what-doesn-t-fit-with

We ignore what doesn’t fit with our biases – even if it costs us #bias #cognition #confirmation[Reposted from 02mysoup-aa]

#regular #snth01

Bias in Criminal Risk Scores Is Mathematically Inevitable, Researchers Say
▻https://www.propublica.org/article/bias-in-criminal-risk-scores-is-mathematically-inevitable-researchers-sa

The racial bias that ProPublica found in a formula used by courts and parole boards to forecast future criminal behavior arises inevitably from the test’s design, according to new research.

The findings were described in scholarly papers published or circulated over the past several months. Taken together, they represent the most far-reaching critique to date of the fairness of algorithms that seek to provide an objective measure of the likelihood a defendant will commit further crimes.

Increasingly, criminal justice officials are using similar risk prediction equations to inform their decisions about bail, sentencing and early release.

The researchers found that the formula, and others like it, have been written in a way that guarantees black defendants will be inaccurately identified as future criminals more often than their white counterparts.

Arrêtez avec les sondages, dit danah boyd chez @hlc
▻http://cfeditions.com/boyd/9novembre/sondages

Il est temps aujourd’hui d’ouvrir un moratoire sur l’usage des #sondages dans les #médias. Et pour les statisticiens d’arrêter de nourrir les enchères médiatiques avec des sondages dont ils connaissent pourtant les #biais.

Les sondages ont une belle histoire, une histoire que la majeure partie de ceux qui se fient aux sondages ne connaissent pas. Dans The averaged American, Sarah Igo décrit trois grandes enquêtes réalisées au milieu du XX siècle qui ont constitué les fondements sur lesquels les instituts de sondage modernes se sont construits : les études menées par Middletown, Gallup et Kinsey. En tant que chercheuse, c’est incroyable de voir la naïveté de ces pionniers, en matière de statistiques et de recueil des données. Mais il y a également un autre constat frappant dans ce livre : bien plus que partout au monde, les Américains avaient envie de contribuer à ce type d’études parce qu’il avaient le sentiment de participer ainsi à la construction de la vie publique. Ils avaient envie de révéler leurs pensées, leurs croyances et leurs idées parce qu’il pensait que cela leur apportait quelque chose, à titre individuel comme à titre collectif.

#bias

Is an algorithm any less racist than a human?
Employers trusting in the impartiality of machines sounds like a good plan to eliminate bias, but data can be just as prejudiced as we are
▻https://www.theguardian.com/technology/2016/aug/03/algorithm-racist-human-employers-work

After firestorm, Facebook training employees to check political biases
▻http://thehill.com/policy/technology/284576-facebook-will-offer-employees-training-to-combat-political-biases

#Facebook will train employees to deal with their political biases, Facebook Chief Operating Officer Sheryl Sandberg said at an event Wednesday evening.

“We think a lot about diversity at Facebook,” she said. “And we have a managing-bias class that all of our leaders and a lot of our employees have taken that I was part of helping to create, and we’ve focused on racial bias, age #bias, gender bias, national bias, and we’re going to add in a scenario now on political bias.”

“So that, as part of [how] we think about helping people understand different points of view and being open to different points of view, we’re dealing with political bias as well going forward.”

Many Silicon Valley companies have turned to unconscious bias training in recent years as they seek to address the tech industry’s lack of diversity in hiring. Among other tactics, they have also started to disclose more about demographics of their workforces.

A spokesperson for Facebook said the training in question is mandatory for all employees.

The decision to include a segment on political biases was sparked by allegations, which Facebook says are unfounded, that editors for the social network’s trending topics feature systematically downplayed news stories and sources popular with conservatives.

Sandberg made the comments Wednesday at an event held by the American Enterprise Institute, a conservative think tank. In addition to the steps Facebook has taken since the bias controversy, she discussed the company’s emerging businesses and “Lean In,” her best-selling book on how women can thrive in the workplace.

Throughout the discussion, she reiterated Facebook’s main argument on many political matters: that it is a neutral platform where users of all political stripes can make their voices heard.

“We don’t have a point of view, we’re not trying to make you have a point of view,” she said.

The Human Side of #Postmortems: Managing #Stress and Cognitive #Bias in #DevOps
▻https://docs.google.com/file/d/0Byl4UKRYLErDVlJMNDNjaThiR2M/edit

The vanishing cost of guessing

▻http://radar.oreilly.com/2013/08/the-vanishing-cost-of-guessing.html

[C]ritics charge that big data will make us stick to constantly optimizing what we already know, rather than thinking out of the box and truly innovating. We’ll rely on machines for evolutionary improvements, rather than revolutionary disruption. An abundance of data means we can find facts to support our preconceived notions, polarizing us politically and dividing us into “filter bubbles” of like-minded intolerance. And it’s easy to mistake correlation for causality, leading us to deny someone medical coverage or refuse them employment because of a pattern over which they have no control, taking us back to the racism and injustice of Apartheid or Redlining.

#données - #prédiction #prévention #anticipation #corrélation - #coût #préjugé

#Daten #Vorhersage #Vorsorge #Analyse #Zusammenhang - #Kosten #Vorurteil - #gelenkte #Wahrnehmung

#big_data #data - #prediction #preventing #analysis - #cost #prejudice #bias #preconception

Apple : « Vous avez 30 heures pour censurer tout ce qui peut paraître pornographique » - Bibliobs avec Le Nouvel Observateur
▻http://bibliobs.nouvelobs.com/web-side-stories/20130404.OBS6844/apple-vous-avez-30-heures-pour-censurer-tout-ce-qui-peut-paraitr

A la fin de la semaine dernière, un responsable d’Izneo a reçu un coup de fil d’Apple. On lui a dit que de nombreuses BD proposées par l’application pouvaient être jugées pornographiques par des Américains. Ce qui ne veut même pas dire qu’elles sont pornographiques. La nuance fait froid dans le dos : à quand la censure d’une photo de femme bras nus sous prétexte qu’un fondamentaliste pourrait trouver ça pornographique ?

Ces #BD, Izneo avait 30 heures pour les trapper, sous peine de disparaître de l’Apple Store. Autant dire disparaître tout court. Mais quelles étaient les BD litigieuses ? « Apple ne nous a donné aucune consigne sur ce qu’ils jugent pornographique ou pas, explique une source anonyme chez Izneo. […] Ils ne nous ont donné aucune clé d’interprétation pour faire le tri. »

La « cellule de crise » a donc fait disparaître toute BD comportant un sein, un décolleté, une courbe de fesse ou une situation sexuellement excitante. Même les très sages Blake et Mortimer en auraient fait les frais, bien qu’on n’ait pas le souvenir du moindre morceau de sein dans les aventures des deux Britanniques

#censure #Apple #pudibonderie #chantage

NewsBusters.org | Exposing Liberal Media Bias
▻http://newsbusters.org
#media #bias #news