#adversarialml search results

Hack The Box released an HTB Academy module on gradient-based evasion (FGSM, I-FGSM, DeepFool). Essential for MSSPs focused on detecting adversarial attacks and hardening client models. Start: okt.to/v70H5W #HackTheBox #MSSP #AdversarialML #ThreatHunting

🙏 Huge thanks to @CienciaGob for supporting this work and to @UAHes for hosting the project. 🌍 If you work on AI security, adversarial ML, or trustworthy AI — let’s connect and collaborate! #TrustworthyAI #AISecurity #AdversarialML #CyberSecurity #MachineLearning

📢 We are excited to share that 4 new papers acknowledging #CoEvolution have been accepted for publication! 🎉 📝 Topics include adversarial robustness, federated learning & 3D perception. 👉 Explore them on our website: coevolution-project.eu/publications/ #AI #AdversarialML #EUFunded

New article! 📖 Discover SecML-Torch, an open-source Python library from @univca’s sAIfer Lab, designed to advance research in Adversarial Machine Learning (AML) and evaluate ML model robustness. 🔗 Read the full article: coevolution-project.eu/secml-torch-a-… #AdversarialML #Cybersecurity

Loved being back in Berlin for @WeAreDevs World Congress. Great to catch up with friends and chat with so many amazing people! 🎉 Recording of my talk “Confuse, Obfuscate, Disrupt” is out now: bit.ly/3I5OEJU #AdversarialML

AI’s greatest strength—its ability to learn—can also be its biggest weakness. Adversarial ML exploits this, turning tiny tweaks into massive threats. From crashing cars to tricking diagnoses, the risks are real. Time to fight back. #AIsecurity #AdversarialML #RobustAI

Adversarial ML is a wake-up call: attackers move fast—so must we. Building resilient AI means secure-by-design systems, constant vigilance, and adapting as threats evolve. The future of AI depends on trust. Let's defend it. #AIsecurity #AdversarialML #TrustworthyAI

NIST finalized a practical playbook #NIST to defend AI 🛡️ with a clear taxonomy of attacks #AdversarialML and an index mapping threats to mitigations for #Cybersecurity teams, giving shared language and faster action #AIsecurity Now the real work is yours as you update threat…

🚨 Adversarial machine learning can deceive autonomous drones ... from GPS spoofing to visual perturbations, slight alterations can trigger misclassification or hijacked navigation, threatening safety and defense systems. Stay alert: defensenews.com/global/europe/… #AdversarialML…

Model evaluation will stop storytellers or scriptwriters to reevaluate or stop crafting adversarial illusions. Metrics don’t lie. Precision exposes the fiction. #AI #ModelEvaluation #AdversarialML #DataEthics

🚨 Deceptive Data Fusion: A Challenge for Real-World ML Ethics 🤖⚖️ 📁 Data_1: Emotionally skewed interaction logs 🧠 Data_2: Behavior spoofing & mimicry models 🔍 Data_3: 3+ years of third-party digital anomalies #AIethics #AdversarialML #DataPoisoning #MLAudit #DigitalRights

6/20 Master adversarial examples. These are inputs designed to fool ML models. A stop sign with strategic stickers can be classified as a speed limit sign. Tools like CleverHans and Foolbox are essential for this. #AdversarialML #ComputerVision

2/20Start with the fundamentals: understand how AI attacks actually work. Adversarial examples aren't just academic - they're real threats. Spend time on FGSM, PGD, and C&W attacks. Get your hands dirty with CleverHans or ART libraries #AdversarialML #AIDefense

Most orgs treat AI as a feature, not an attack surface. VerSprite’s new guide on AI Red Teaming shows how to simulate real-world attacks on ML pipelines, LLMs, and model supply chains. Read: versprite.com/blog/ai-red-te… #AIsecurity #RedTeam #AdversarialML

AI is transforming our world—but are we building it responsibly? From adversarial attacks to ethical dilemmas, the future of AI depends on how we integrate, defend, and govern it. #AI #ResponsibleAI #AdversarialML #Cybersecurity #TechPolicy

Los ataques adversarios de IA no solo hackean sistemas: cambian su percepción de la realidad. ➠ Sistemas de crédito con sesgos ocultos ➠ Drones engañados por patrones invisibles ➠ Modelos de reconocimiento facial robados #MachineLearning #AdversarialML buff.ly/8ZbtYmG

Los ataques adversarios de IA no solo hackean sistemas: cambian su percepción de la realidad. ➠ Sistemas de crédito con sesgos ocultos ➠ Drones engañados por patrones invisibles ➠ Modelos de reconocimiento facial robados #MachineLearning #AdversarialML buff.ly/8ZbtYmG

Adversarial AI attacks don’t just hack systems—they rewrite how AI sees reality. ➠ Poisoned loan algorithms that discriminate ➠ Drones fooled by invisible patterns ➠ Stolen facial recognition models Battlefield is math. #AdversarialML #MachineLearning buff.ly/8ZbtYmG

Adversarial AI attacks don’t just hack systems—they rewrite how AI sees reality. ➠ Poisoned loan algorithms that discriminate ➠ Drones fooled by invisible patterns ➠ Stolen facial recognition models Battlefield is math. #AdversarialML #MachineLearning buff.ly/8ZbtYmG

Full P2 room for the #adversarialML tutorial at #DeepLearn2019 @PRA_Lab @pluribus_one #deeplearning #NeuralNetworks #MachineLearning

"Order-Disorder: Imitation Adversarial Attacks for Black-box Neural Ranking Models", 1st April, led by @VictorKnox99 #CCS2022 #AdversarialML [9/N]

@alidehghantanha speaking on #AI and #adversarialML and running #malware analysis workshop at the International Summer School on #ComputationalForensics (#SuCoFo2019)

[Adversarial ML] Learn and understand about Adversarial Machine Learning and Adversarial Threat Models. Link : blog.cyberwarfare.live/adversarial-at… Website : cyberwarfare.live Email : [email protected] #redteam #adversarialML #offensivemachinelearning #machinelearning

Nos vemos a las 6pm (hora Colombia) para hablar de seguridad en inteligencia artificial en el @DragonJARCon @icesi @Editorial_Icesi #MachineLearning #DataScience #adversarialml #BigData #infosec #hacker

To identify parts of the sentence vulnerable to attack, we create a constituency parse tree and use a perplexity-difference based metric to gauge the phrase peculiarity at each node. We prioritise our attack on the top-N nodes with the most peculiar phrases. #AdversarialML [4/N]

Awesome #adversarialml research at poster session today at @icmlconf

Hack The Box released an HTB Academy module on gradient-based evasion (FGSM, I-FGSM, DeepFool). Essential for MSSPs focused on detecting adversarial attacks and hardening client models. Start: okt.to/v70H5W #HackTheBox #MSSP #AdversarialML #ThreatHunting

I asked blenderbot.ai if they would be aware of someone intentionally attacking them (or not) … interestingly they now think I’m both a student and a liar. One of those is true! #adversarialML #cybersecurity

Machine Learning Attack Series: Image Scaling Attacks embracethered.com/blog/posts/202… Hiding an image inside another. Hidden image becomes visible when server rescales it. 🤯 #machinelearning #adversarialml #offensiveml #redteaming #pentest #infosec #ml #aiml

We can live with a panda recognized as gibbon, but what if a stop sign is made invisible to computer vision algorithms by adding small stickers? #adversarialML

Great technical talks this morning from speakers at @Facebook, @McAfee and @unboxresearch exploring the important steps we can take to building responsible, inclusive AI systems. #EGGSF2019 #datascience #adversarialML #MachineLearning #ML

(3rd chapter) Different types of Adversarial ML Attacks: - Perturbation - Membership inferencing - Model stealing - Deep NN Jailbreaking - Physical domain poisoning - Training Data Reconstruction - Poisoning attack - Model backdoor attack #machinelearning #adversarialML

(1/3) Adversarial Machine Learning safeguards #AI models against attacks. By understanding vulnerabilities and manipulating data inputs, robust defenses are built to ensure system integrity, reliability, and security. #TechSimplified #AdversarialML #AIsecurity

I was really excited and honored to give a talk about #tensors for #graphmining and #adversarialML at the One World Signal Processing seminar series yesterday! Special thanks to the organizers @XOsueecs @yanning_shen @hoitowai!!! #graphs #machinelearning #signalprocessing

🚨 Adversarial Machine Learning: quando l'IA viene ingannata! 🧠💻 . . . #AdversarialML #MachineLearning #AI #CyberSecurity #adversarialmachinelearning #sicurezza #ricerca #etica #informatica #massaecozzile #pistoia #montecatiniterme #consulenteinformatico #studioinformaticodg

How does an approximate multiplier design act as a defense against adversarial attacks? Defensive Approximation: Securing CNNs using Approximate Computing asplos-conference.org/abstracts/aspl… @AmiraGuesmi4 @ihstein @nael_ag #ASPLOS21 #AdversarialML #Approximation

Slides for "The History of Adversarial AI" are online! See the quick summary of the past 10 years in AI vulnerability research. Get PDF from the #HITB2021AMS website: conference.hitb.org/hitbsecconf202… #AdversarialML #SecureAI #TrustworthyAI #ResponsibleAI

@Bushra_Sabir led an insightful discussion on "#AdversarialML in #Cybersecurity" #crest_discussion. We discussed the key #whitebox, #greybox & #blackbox attacks & defense mechanisms in various #AI-enabled #security tasks such as #malware, #phishing, #spam, #intrusion detection.

Yep, our adversarial defense paper is accepted #ECCV! Context modelling is used to detect adversarial examples out of context, e.g., a speed limit sign (should be a stop sign) at a crossing road with stop line. Preprint and code coming soon.#adversarialML @laosong @zst_rising88