Discover and read the best of Twitter Threads about #algorithm

Most recents (16)

So it turns out that not only is @zoom_us' #FacialRecognition #algorithm so #racist it chops black people's heads off...

..but @Twitter's image processing #algorithms are #racist too:



It's getting to the point where we need an industry reboot; if #SiliconValley genuinely wants to build '#EthicslAI', it must delete 'em all and start again. Respecting #HumanRights this time.
Read 4 tweets
@myESMO #ESMO20 as a #trainee can be #overwhelming! So many good studies, some more #practicechanging then others, if you missed some and want to understand (albeit at a simplistic #trainee level), sit back, relax and enjoy as we go through some great data #ESMO20 @OncoAlert
1. #NSCLC: 2 major studies #ADAURA #CROWN for adjuvant #EGFRmNSCLC, and advanced #ALK+ experts can provide better perspective @JackWestMD @n8pennell @StephenVLiu @AMansfieldMD @CharuAggarwalMD @NarjustDumaMD @GlopesMd @DevikaDasMD @OncoAlert
1. A) #ADAURA: Stage IB-IIIA #resected #NSCLC with #EGFRm treated with #Osimertinib vs #placebo [SOC prior to this was adjuvant chemotherapy [cisplatin-based doublet based on #LACE metanalysis- pubmed.ncbi.nlm.nih.gov/18506026/] showed improvement in #DFS @NEJM nejm.org/doi/full/10.10…
Read 19 tweets
So @cabinetofficeuk (@GDSTeam) has updated the #DataEthicsFramework for government and the public sector today - the first version of which was published by @MattHancock in 2018, when he was at @DCMS:
gov.uk/government/pub…

Two obvious questions:

1) Where are the published...
...ratings for all of the government & public sector data projects / programmes that have been started since 2018?

A framework that neither demands nor provides publicly visible outputs surely fails on two of its own 'overarching principles'; #transparency and #accountability...
2) Other than "consult with your team leader", what happens when a project 'fails' the framework - say a national programme that doesn't do a #DPIA, or a body that employs a #discriminatory #algorithm?

An 'ethics' framework without meaningful #consequences ain't #ethical at all.
Read 6 tweets
... it's best to stop digging:

theguardian.com/education/2020…

(Waving shovels around can be a dangerous business...)
The letter @RTaylorOpenData sent to @RoyalStatSoc last Friday (21Aug):
assets.publishing.service.gov.uk/government/upl… selectively quotes from what @RoyalStatSoc said to @StatsRegulation: Image
Read 8 tweets
So at exactly what point "over the weekend" did @GavinWilliamson realise the system for which he is responsible, which should have been checked thoroughly, was not "fair and robust"?

thetimes.co.uk/article/gavin-…

The Minister (who's clearly fond of his whips) is far too cavalier. Image
This ain't over by a LONG way, and while there are no right answers right now - cf. this and rest of [Thread]: - @10DowningStreet would do well to reassure (future) voters it has people with a proper sense and appropriate grasp of detail in place...
...#data, #AI and #algorithm-obsessed Governments that avoid #accountability and treat their own citizens as just so much '#GarbageInGarbageOut' don't have much of a future.

Mark. My. Words!
Read 3 tweets
Day 1: Horror and anguish
Day 2: Ministers double down
Day 3: "Fuck the #algorithm!"
Day 4: Government U-turn 👇
theguardian.com/education/2020… Image
(Up in Wilmslow / over in Canada #ICO breathes a sigh of relief at a dodged bullet...)
Though my thoughts go out to university admissions staff, who must be bulk-ordering ProPlus and paracetamol right now...
Read 3 tweets
"The #OFQUAL #algorithm is a gross misuse of statistics... there is a special place in hell for those who use maths to hide rather than reveal the truth, this algorithm is hiding the achievement of this generation of young people." my @ChronicleLive blog chroniclelive.co.uk/news/news-opin…
Most people now think the #AlevelResults should not stand, whether the Govt will recognise that or not is another matter. But we shouldn't blame #algorithms for this huge injustice - like most technology they are neither good nor bad in themselves
And I don't blame those who wrote the #algorithm I do think #engineers (like me) & #datascientists should take a 'do no harm' oath like doctors, but currently there is no framework for that & they were presumably working to a spec
Read 6 tweets
In the 00s we phrased it, "Stop the #DatabaseState".

I'm doubly glad to see this: firstly because it's great to see the next generation pick up the refrain. Secondly 'cos they're applying the appropriate level of profanity!
By the way, if you want to know who coined the phrase #DatabaseState, it was @guy_herbert (in the early 'id-core' discussions of @NO2ID in 2004).

Here he is today:


And for a list of shit we helped stop, read Section 3 of this:
assets.publishing.service.gov.uk/government/upl…
That some of the very same politicians who professed to oppose the #DatabaseState back in 2010 so blindly now worship the #algorithm shows how little they ever actually #understood. Also how little they can be #trusted...

WTF else do they think "data-driven Government" means?
Read 3 tweets
Maybe not as prominently, but it's been with us for a long time: from the application of #probability to testimony & judgment from late 1700s, to #RacialDiscrimination by the actuaries of the late 1800s, to the eugenic evil of #IBMandTheHolocaust. That we're now more conscious...
...of 'the #algorithm' doesn't mean we've become any better at preventing its abuse.

#AppliedStatistics is still inherently #amoral and more often than not #unethical-by-default now it's at #AI speed, and wielded by those who value #prediction over human rights & human lives...
...and the question remains, now public awareness is rising: what will we DO about it this time?

Reactive outrage, protest and #wackamole litigation (as necessary as those may be) will not deliver the strategic, systemic solutions we need for a just and fair #InformationSociety.
Read 3 tweets
Good luck & best wishes to all those receiving their #AlevelResults & #vocational results today in #Newcastle & across the country. You have faced challenges no other year group has had to endure. You are to be congratulated in everything you achieve.
Hearing VERY WORRYING examples of #Newcastle #AlevelResults slashed by nationally imposed #algorithm that does not reflect individual achievement. Constituents who have suffered unfair grade cuts, get in touch. We can't allow young people's & our city's future to be so damaged.
I understand that in particular girls studying #Physics have been impacted by an algorithm which does not take into account all the work by schools (& campaigners!) to increase & widen girls participation in Physics, some losing places to study #Engineering #AlevelResults
Read 5 tweets
We already do. And not just in #health.

Matthew mentions the #QALY ('Quality-Adjusted Life Year') that, via #NICE, is currently valued at between £20-30k - but QALYs are used, and valued differently, across different Gov't Departments too: qna.files.parliament.uk/qna-attachment…
Meanwhile, 'in the private sector', #insurance payouts for #death remain significantly lower than for debilitating #injury, cf. #JudicialCollegeGuidelines: quittance.co.uk/personal-injur… and legislation like the Fatal Accidents Act 1976: legislation.gov.uk/ukpga/1976/30/…

So maybe, rather than...
...haggle, posture and virtue signal over numbers on which no-one will ever agree, but which #actuaries, #algorithm makers and #AI trainers WILL inevitably use, we should focus on far more meaningful debates - like what we want '#SocialCare' to actually mean?
Read 3 tweets
Daily Bookmarks to GAVNet 6/06/2020-2

greeneracresvaluenetwork.wordpress.com/2020/06/06/dai…

Coronavirus May Be a Vascular Disease, Which Explains Everything | Elemental

elemental.medium.com/coronavirus-ma…

#vessels #blood #coronavirus #disease
Out-of-Sync ‘Loners’ May Secretly Protect Orderly Swarms

abstractions.nautil.us/article/561/ou…

#individual #swarm #nautilus
Opinion | The ‘Liberal World Order’ Was Built With Blood - The New York Times

nytimes.com/2020/05/29/opi…

#liberal #blood #order #world
Read 12 tweets
Daily Bookmarks to GAVNet 05/11/2020 -2

Early Herd Immunity against COVID-19: A Dangerous Misconception - Johns Hopkins Coronavirus Resource Center

coronavirus.jhu.edu/from-our-exper…
#HerdImmunity
Robot reminds visitors of safe distancing measures in Bishan-Ang Mo Kio Park, Singapore News & Top Stories - The Straits Times

straitstimes.com/singapore/robo…
#Singapore #Distance #dystopian #Robot
A Much-Hyped COVID-19 Drug Is Almost Identical to a Black-Market Cat Cure

theatlantic.com/science/archiv…
#China #drugs #blackmarket #coronavirus
Read 8 tweets
What is #dataracism?

See how data-driven technologies target & disproportionately impact on racialised people at the margins: enar-eu.org/Data-racism-a-… #AI #algorithms

@EU_Justice @EURightsAgency @PICUM_post @MPriv_o
@netzpolitik_org @RebekahDelsol @laquadrature @sarahchander Image
@EU_Justice @EURightsAgency @PICUM_post @MPriv_o @netzpolitik_org @RebekahDelsol @laquadrature @sarahchander @e3i5 @BrusselsGeek @AlexandraGeese @edri @InterwebzNani @samiraraf @DataJusticeLab @OSFJustice @privacyint @StopWatchUK Die Schattenseite der „Innovation“ und der verbesserten „Effizienz“ automatisierter Technologien liegt darin, wie sie marginalisierte Gruppen – rassifizierte Menschen, queere Communities und Menschen mit Behinderungen – kategorisieren, #AI #Dataracism
netzpolitik.org/2020/eine-neue…
Halt public use of facial recognition tech! says @EHRC.😠

Mass screening of the public at shopping centres or events like pop concerts, by police using #FacialRecognition, must be halted: theguardian.com/uk-news/2020/m…

#Datenrassismus #AI #algorithm #dataracism @MPriv_o @JuliePascoet Image
Read 3 tweets
The gender biased #AppleCard credit decision is a fascinating example of how algorithmic credit decisions can enter choppy waters
Even more fascinating is Goldman’s response. They can’t identify the reason there was such a difference in credit limits but they are sure that decisions are not based on “factors like gender”
In fact, they have no way of guaranteeing that. The fact they exclude protected characteristics or “do not know your gender or marital status” does not mean these factors were not used. These are #AIFalseDreams
Read 7 tweets
“We have this idea that #machines can liberate us from repetitive tasks, but there are different templates to use this #technology...” @PauloBlikstein #AIED2018 @AIEDconf | @uclknowledgelab @EDUCATEldn @AIConf2018 @brussels_ai @digityser @results4dev @CEInnovations @audreywatters
“Another element in this template, is that #learning happens at different paces... it changes from person to person ... but the awareness of the limitations has somehow been lost over the decades” @PauloBlikstein #AIed @AIEDconf #personalizedlearning @uclknowledgelab @EDUCATEldn
But...here’s the problem

“There is a big problem with this template that we have been using for decades ... in using this map...” @PauloBlikstein

#AIed18 #ISTE18 #EdMedia18 #DISummit @AIEDconf @uclknowledgelab @EDUCATEldn @EdSurge #Edtech @PodcastEdtech @EdtechukHQ
Read 5 tweets

Related hashtags

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!