Federated Learning: Revolutionizing Machine Learning with Privacy-Preserving Techniques
Federated Learning (FL) is redefining how machine learning models are trained, allowing for decentralized data processing while safeguarding user privacy. This innovative approach not only enhances da...
This article does a great job highlighting why DOGE is taking over the federal government so easily: federal systems centralize massive amounts of sensitive data, making them highly vulnerable to insider threats. The article concludes by pointing out that techniques like #FederatedLearning and #DifferentialPrivacy could help build more resilient systems 
#Privacy #CyberSecurity
Der Sammelband ist jetzt als Open-Access-Publikation im Nomos Verlag erschienen.
Jetzt kostenlos als eBook lesen oder als Print bestellen!
https://researchinstitute.at/neu-beitrag-zu-virtuellen-zwillingen-und-diabetes-open-access/
#introduction #MachineLearning #TrustworthyAI #decentralization #privacy #decentralized #FederatedLearning #DifferentialPrivacy #data #AI #Fediverse
I am really interested in these topics and I would like to hear every now and then some news or paper recommendation or sth like that.
Time for a proper #introduction! 
I'm a #MachineLearning researcher at Inria, France, focusing on #TrustworthyAI—especially #decentralization & #privacy. I design algorithms that learn from #decentralized data (#FederatedLearning) without memorizing personal data (#DifferentialPrivacy), working toward putting #data & #AI back in people's hands. Maybe one day, these ideas could be used in the #Fediverse!
"Approximate and Weighted Data Reconstruction Attack in Federated Learning" co-authored by #MSCA researcher Ziqi Wang
demonstrates that private data can be reconstructed from shared information in federated learning emphasizing the urgent need for stronger defense mechanisms to protect user privacy.
#federatedLearning
#gradientInversion
> HIS story of #satireSpotlights
What do you mean @erio7,
the #realSatirePur moments in @jesuisatire's #fediLife?
Like, HIS responsibility for the first urgent over night role out of a blocking feature (seriously!) in the #fediVerse, because HE tried to have fun with a troll who didn't had enough #selbstIronie in his ego, nor respect for the decentralized system or the community, showing of in the worst irresponsible manner possible?
Or the #meanwhile incident, when "diaspora.com.ar … on it’s quest for the #holyTroll meet jesuiSatire .. ᘛ⁐̤ᕐᐷ" and called for the first time in #theFederatedHistory for a instance wide blocking of a distant user profile, wondering how to implement that in the code, telling HIM to quote: "¡stop being ironic!", with the admin writing an extense howTo defend yourself from a troll opEd, while HE already was trying to snip away users by writing takes like:
#federatedlearning, #hola #Argentina @Elcammarota
"About how to interpret a profile in the internet and why it can be useful to give #whiteHatTroll's ( as described by @linuzifer ) a chance."
¿ 
?
HIS april fools censorship! prank conversation that hit the cord?
Or the moment HE upset @lizzischmidt with HIS #opEd in Englisch promoting her German #sundayGimp movement by stating "they even are arrogant enough to invent new hashtags" and she took it serious?
That post nearly ended a friendship and the whole community had to come to the rescue to translate and state the obvious.
Like to say:
"Even @deusfigendi kann das bezeugen".
We wonder if you are aware that you are talking about THE ONE who wrote the first draft of the #TSGDC more than 13 years ago, a take that actually could need some refurbishment, spellcheck and renaming to #TSGFC (The Sporial Guide to the Fediverse Cyberspace).
Well, actually we think HE is best when HE shows of with HIS 144 character bilingual or trilingual tweete takes like:
"German for #newBees leasons, because they have a word for everything."
(Ok, that one was really what he is all about, kinda selbstentlarvend. In other words, for a member of the #MUH a little bit to dangerous perhaps)
Anyway, please don't make us remember old wounds, scares and stripes, we came a long way already ..
.. and be assured, @requeteChe was right when he stated from the get go:
ʕ•ﻌ•ʔ jeSuisatire, sounds like a typical bastard!
The first #ACSAC2024 #PaperPreview today is Li et al.'s FedCAP, is a robust FL framework against both data heterogeneity and Byzantine attacks through customized aggregation and personalization:
https://www.openconf.org/acsac2024/modules/request.php?module=oc_program&action=summary.php&id=336
#AI #FederatedLearning #Cybersecurity
Federated Learning: A world where your data is safe yet insightful! 
Discover how this AI revolution boosts privacy & accuracy in various industries. Follow for more! 
#FederatedLearning #DataPrivacy https://t.co/KyJeLQk5bf https://medium.com/@falitroke/unlocking-the-future-with-federated-learning-f2df219cb525
Flower: A Friendly Federated Learning Framework 
Unified federated learning, analytics, and evaluation. Federate any workload, ML framework, and programming language.
#FederatedLearning #MachineLearning #DataScience #AI #Analytics #Evaluation #TechInnovation #FlowerFramework
Flower: A Friendly Federated L...
Flower: A Friendly Federated Learning Framework
Unified federated learning, analytics, and evaluation. Federate any workload, ML framework, and programming language.
The final paper in this session was Castillo et al.'s "FLEDGE: Ledger-based Federated Learning Resilient to Inference and Backdoor Attacks" which presented a novel approach to enhance #FederatedLearning #Security. (https://www.acsac.org/2023/program/final/s265.html) 5/5
The third paper in this session was Cui & Mei's "ABFL: A Blockchain-enabled Robust Framework for Secure and Trustworthy Federated Learning" which discusses a resilient architecture combining #blockchain and #FederatedLearning. (https://www.acsac.org/2023/program/final/s133.html) 4/5
#FL #FederatedLearning #Data #Ethics "Although FL removes the option of stealing the data from a central database, it does not give any formal guarantees against re-identification. Numerous studies have shown that it is possible to reverse engineer privacy-sensitive data from the model information that is shared in FL. Thus, the term ‘privacy-preserving’ is not entirely appropriate for FL." #privacy #federatedlearning #AIEthics #LLM https://www.nature.com/articles/s42256-024-00813-x
Herzlichen Glückwunsch! Für seine Arbeit im Bereich #DigitalHealth hat Bjarne Pfitzner auf der Fachkonferenz #AIMHC in der Nähe von Los Angeles den Best Student Paper Award erhalten. Bjarne ist Doktorand am Fachgebiet "DigitalHealth – Connected Healthcare" am HPI.
In seinem Paper, das in Kollaboration mit Partnern an der @ChariteBerlin entstanden ist, beschäftigt er sich mit #FederatedLearning in der #Medizin.
Was er genau erforscht hat, erfahrt ihr unter: https://hpi.de/news/jahrgaenge/2024/best-student-paper-award-fuer-bjarne-pfitzner.html
Project announcement 
: "PROSurvival" builds a collaborative federated learning framework to predict survival in prostate cancer patients.
Partners are OFFIS - Institute for Information Technology, Charité, Goethe University Frankfurt, and Fraunhofer MEVIS.
Here's the gist: In the long run, we want to find predictive image features that can be identified in tissue sections from routine diagnostics.
#federatedlearning #ai #foundationmodels #ComputationalPathology
1/3
Freuen Sie sich: Dr. Daniel Gille von der @Cyberagentur auf dem Online-Symposium „Federated Learning & ML Security“ am 27.03.24! Sein Thema "Machine Learning & IT-Security – Forschungsbedarfe für die Innere und Äußere Sicherheit".
Spannende Diskussion beim #MANTRASymposium: Wie können wir die #ITSicherheit von #FederatedLearning verbessern?
Anmelden: #FLMLSec2024: https://t1p.de/bipwu
Mehr: https://t1p.de/rit9d
#Lösungsansätze #KI #ML #Cybersicherheit #KI #ML #ITSicherheit #DataScience
“pfl is a Python framework developed at #Apple to empower researchers to run efficient #simulations with #privacy-preserving #federatedLearning (FL) and disseminate the results of their research in FL. We are a team comprising engineering and research expertise, and we encourage researchers to publish their papers, with this code, with confidence.”
MicrobAIome verwendet KI und Mikrobiomdaten, um die Behandlung von Darmkrebs zu verbessern, ohne dabei die Privatsphäre der Patient*innen zu gefährden. Es nutzt föderiertes Machinelearning und integriert die Ergebnisse der Untersuchungen nationaler Datenbanken in ein sicheres, internationales Netzwerk, um evidenzbasierte Richtlinien für die Behandlung zu entwickeln.
Mehr dazu findest du unter: https://researchinstitute.at/forschung/#8bcaf51ea3197d59e
Vom 7. bis 8. Februar versammelte sich das Team von #MicrobAIome in Krakau, Polen, zur Generalversammlung 2024.
Alle Projektpartner aus Irland, Frankreich, Rumänien, Österreich und Deutschland waren unter einem Dach vertreten und vereint, um sich in Diskussionen über Projektfortschritte einzubringen und den Weg in die Zukunft zu gestalten!
