Create accountLogin

Recent searches

No recent searches

Search options

Only available when logged in.
shakedown.social is one of the many independent Mastodon servers you can use to participate in the fediverse.
A community for live music fans with roots in the jam scene. Shakedown Social is run by a team of volunteers (led by @clifff and @sethadam1) and funded by donations.

Administered by:

Server stats:

292
active users

shakedown.social: AboutStatusProfiles directoryPrivacy policy

Mastodon: AboutGet the appKeyboard shortcutsView source codev4.3.4

#machinelearning

14 posts12 participants0 posts today
Public *
Daniele de Rigo

1/

In 2018, on March 17 the so called "#CambridgeAnalytica Files" were published by the Observer (Guardian Media Group, web.archive.org/web/2018031801).

The journalistic investigation by Carole Cadwalladr and Emma Graham-Harrison [1] discussed one of the first examples of digital social engineering in the context of potentially subverting democratic processes via #SocialMedia and #MachineLearning techniques with custom #manipulation of each accessible individual elector.

Overview in the video [2].

the GuardianThe Cambridge Analytica Files | News | The Guardian
Continued thread
Public
Mx Verda

Also: can we have an AI Machine Learning LLM whatever thing trained ONLY on consentually offered materials, text and art?
The bare minimum effort of "you must send your materials to us" rather than trawling the internet to steal whatever isn't nailed down.
Sure, spam bots and nuisance or malicious submitters, but most won't be. Could always have a holding pen area of like 2 or 3 months before adding to training data.

Proactively request opt-outs as well, to make filtering a tiny bit easier?

#AI#OpenAI#ChatGPT
Public
Osma Suominen

The German National Library @DNB_Aktuelles AI team just published a paper "Automatic Subject Cataloguing at the German National Library" in LIBER Quarterly (open access):

doi.org/10.53377/lq.19422

The paper documents in detail how they use open source tools such as Git, Data Version Control and Annif to implement automatic assignment of DDC classes and GND subject terms to digital publications in various formats. Great work!

doi.orgAutomatic Subject Cataloguing at the German National Library | LIBER Quarterly: The Journal of the Association of European Research Libraries
#libraries#cataloguing#opensource
Replied in thread
Public *
doboprobodyne

@nullagent
Of course the best answer (for a given definition of best) is to avoid the region and make it as unprofitable as practicable, while finding ways to support conventional policing and strengthen democratic institutions in other regions.

So it is therefore clearly on a completely unrelated note that I share this link to amusing printed masks of your favourite topical celebrities and politicians, (from a Chinese supplier, please do share a link to manufacturers in friendlier states, if anyone has one).

aliexpress.com/item/1005008613

🤪

aliexpress.Funny 3D Printed Realistic Face Mask Elastic Mesh Full Face Costume Mask for Men Women Cosplay Balaclava Hood Party Cosplay Prop - AliExpress 15Smarter Shopping, Better Living! Aliexpress.com
#comedy#surveillance#policing
Public
Miguel Afonso Caetano

"There is an old maxim that ‘every model is wrong, but some models are useful’. It takes a lot of work to translate outputs from models to claims about the world. The toolbox of machine learning makes it easier to build models, but it doesn’t necessarily make it easier to extract knowledge about the world, and might well make it harder. As a result, we run the risk of producing more but understanding less.

Science is not merely a collection of facts or findings. Actual scientific progress happens through theories, which explain a collection of findings, and paradigms, which are conceptual tools for understanding and investigating a domain. As we move from findings to theories to paradigms, things get more abstract, broader and less amenable to automation. We suspect that the rapid proliferation of scientific findings based on AI has not accelerated — and might even have inhibited — these higher levels of progress."

nature.com/articles/d41586-025

www.nature.comWhy an overreliance on AI-driven modelling is bad for scienceWithout clear protocols to catch errors, artificial intelligence’s growing role in science could do more harm than good.
#AI#Science#ML
ExploreLive feeds
About