Mastodawn

Adversarial Attacks: When Machines Are Fooled by the Invisible
https://youtu.be/cpm7I0KX_kA #CyberSecurity #ArtificialIntelligence #AdversarialAI #MachineLearningSecurity #AISecurity #AdversarialExamples #AIThreats #CyberRisk #ResponsibleAI #AutonomousSystems #FraudDetection

News Beep Aug 13, 2025

A new way to test how well AI systems classify text | MIT News

Is this movie review a rave or a pan? Is this news story about business or technology? Is…
#NewsBeep #News #Headlines #adversarialexamples #chatbots #KalyanVeeramachaneni #largelanguagemodels(LLMs) #Latvia #LeiXu #LV #MITIDSS #MITLIDS #SarahAlnegheimish #SP-Attack #SP-Defense #syntheticexamples #textclassifiers
https://www.newsbeep.com/58541/

Show thread

it's B! Cavello 🐝Mar 14, 2024

This also has me thinking about SEO hacks through the lens of #AdversarialExamples. Something that’s interesting with natural language (vs, say, least significant bit steganography or pixel-level changes on adversarial images) is that the “resolution” is very low, meaning human perception of difference tends to be higher.

Martin Gubri Jun 21, 2023

Live PhD Defence of Martin Gubri "What Matters in Model Training to Transfer Adversarial Examples"

https://aperi.tube/videos/watch/8bc3bdce-497b-4a17-8cb1-14994cef723a

Live PhD Defence of Martin Gubri "What Matters in Model Training to Transfer Adversarial Examples"

PeerTube

🐦‍🔥nemo™🐦‍⬛ 🇺🇦🍉Feb 5, 2023

"Deep dive into the world of #AdversarialExamples! Learn how they can fool #NeuralNetworks & impact real-world #MachineLearning applications. Join us on a journey to understand why they exist & what can be done to prevent them. #FoolingNeuralNetworks #MLSecurity #Reliability"

https://nemozone.writeas.com/fooling-neural-networks-a-deep-dive-into-adversarial-examples

Fooling Neural Networks: A Deep Dive into Adversarial Examples

In the world of machine learning, neural networks are widely used for a variety of tasks, including image recognition, natural language p...

nemozone

Quincy ⁂Apr 20, 2019

So. Mein Avatar-Image ist nun ein Sturmgewehr!!! (*)

Jedenfalls nach dem Urteil von Inception V3 + ImageNet: Ergebnis meiner ersten Bastelstunde mit Tensorflow. #AdversarialExamples leicht gemacht dank exzellentem interaktivem Tutorial von Anish Athalye https://www.anishathalye.com/2017/07/25/synthesizing-adversarial-examples/

Los, Terrorfilter, zeig, was du kannst!

(*) Ob das eine gute Idee ist? Keine Ahnung. Aber vielleicht sollte man sich eher fragen, ob es eine gute Idee ist, angebliche "künstliche Intelligenz" überall einzubauen ...

A Step-by-Step Guide to Synthesizing Adversarial Examples

Synthesizing adversarial examples for neural networks is surprisingly easy.

feed Sep 27, 2018

Hate-speech detection algorithms are trivial to fool https://boingboing.net/2018/09/27/ha-ha-only-serious.html #adversarialexamples #platformcensorship #computerscience #machinelearning #happymutants #scholarship #security #bigtech #cda230 #Post #AI #ml

Hate-speech detection algorithms are trivial to fool

In All You Need is “Love”: Evading Hate Speech Detection, a Finnish-Italian computer science research team describe their research on evading hate-speech detection algorithms; their work will be pr…

Boing Boing

feed Sep 25, 2018

There's a literal elephant in machine learning's room https://boingboing.net/2018/09/25/pachydermal-adversarial.html #adversarialexamples #imageclassifiers #computerscience #machinelearning #happymutants #scholarship #Post #AI #ml

There's a literal elephant in machine learning's room

Boing Boing

Boing Boing Mar 27, 2018

Law professors and computer scientists mull whether America's overbroad "hacking" laws ban tricking robots https://boingboing.net/2018/03/27/danger-will-robinson.html #everythinglookslikeanail #adversarialexamples #computerscience #machinelearning #scholarship #Post #cfaa #law #AI

Law professors and computer scientists mull whether America's overbroad "hacking" laws ban tricking robots

Boing Boing

Boing Boing Mar 26, 2018

Invisible, targeted infrared light can fool facial recognition software into thinking anyone is anyone else https://boingboing.net/2018/03/26/the-threaten-from-infrared.html #adversarialexamples #computerscience #facerecognition #machinelearning #authentication #scholarship #security #stupidai #infosec #china #Post #AI

Invisible, targeted infrared light can fool facial recognition software into thinking anyone is anyone else

A group of Chinese computer scientists from academia and industry have published a paper documenting a tool for fooling facial recognition software by shining hat-brim-mounted infrared LEDs on the …

Boing Boing