SecMLOps -arxiv.org/pdf/2601.10848 Integrating Security Throughout the Machine Learning Operations Lifecycle Secure Machine Learning Operations (SecMLOps), providing a comprehensive framework designed to integrate robust security measures throughout the entire ML operations (MLOps) lifecycle. SecMLOps builds on the principles of MLOps by embedding security considerations from the initial design phase through to deployment and continuous monitoring. This framework is particularly focused on safeguarding against sophisticated attacks that target various stages of the MLOps lifecycle, thereby enhancing the resilience and trustworthiness of ML applications. With the increasing concerns over ML security risks, the concept of Secure Machine Learning Operations (SecMLOps) was proposed to extend the MLOps with security considerations. This paradigm advocates for the explicit integration of security measures throughout the entire MLOps lifecycle. By embedding security considerations from the outset, SecMLOps aims to cultivate more secure, reliable, and trustworthy ML-based systems. This holistic security integration not only enhances the resilience of ML deployments but also ensures their alignment with organizational security policies and regulatory requirements, thereby fortifying the foundation of trust and dependability in ML applications across various sectors. Xinrui Zhang, Pincan Zhao, @JasonJaskolka, @henglli, Rongxing Lu - @Carleton_U, @polymtl, @queensu #SecMLOps #MLOps #MachineLearningSecurity #AdversarialMachineLearning #AdversarialExamples #DataPoisoning #STRIDE #ThreatModeling #AdversarialTraining #ModelServing #CityPersons #PedestrianDetection

🧠💻 ¿Se puede hackear una IA? Claro que sí… y en The Hackers Labs lo estamos haciendo. linkedin.com/posts/the-hacke… #HackingIA #HackearModelos #TheHackersLabs #AIHacking #Ciberseguridad #MachineLearningSecurity #AdversarialExamples #CTF #EthicalHacking #RedTeamIA #DeepLearning

実は、AIが「考えている」時の脳内（ニューラルネット）活動は、人間の直感や思い込みに近いことがある。 たとえば、ある画像を「ゾウ」と判断するAIモデルに、ノイズを1ピクセル加えるだけで「バナナ」と認識することがある。 これは「敵対的サンプル（adversarial example）」と呼ばれ、AIの“直感”の脆弱性を突く技術。 面白いのは、これが人間の錯覚に似ている点。 自分達が2本の平行線を「曲がって見える」と錯覚するように、AIも訓練の偏りやデータ分布の罠にハマる。 つまり、AIも“見たいもの”を見ている。まるで夢を見ているように。 #AI豆知識 #機械学習 #人工知能 #AdversarialExamples

Research and development of state-of-the-art deepfake detection analytics with intuitive explanations and robustness to open-world variations as well as malicious adversarial examples. #adversarialexamples #deepfakedetection #robustai

Ecco il video del seminario che ho fatto a @unimib su #adversarialexamples e #differentialprivacy. Con questo inauguro il mio canale, e quindi: iscrivetevi e lasciate like, che altri contenuti interessanti stanno per arrivare. 😉 youtu.be/IDwHzE44MAA via @YouTube

Mercoledì alle 11 darò un seminario al dipartimento di informatica di @unimib su #AdversarialExamples nei modelli di #DeepLearning, e come contrastarli con la #DifferentialPrivacy. Dettagli nella locandina. Se siete in zona, siete benvenuti! Il seminario sarà anche registrato.

• قد تحدث هلوسة الذكاء الاصطناعي بسبب الأمثلة العدائية أو فك ترميز المحولات غير الصحيحة في النماذج الكبيرة القائمة على اللغة مثل ChatGPT.  #AHallucination #AdversarialExamples

Unsere Wissenschaftler zeigen heute auf der AI.BAY 2023 ihre neusten Forschungsergebnisse zum Schutz von #KI vor Manipulation & Angriffen. Jetzt kostenlos dazuschalten: aisec.fraunhofer.de/de/press… #Deepfake #AdversarialExamples #WeKnowCybersecurity

#KI ist aus unserem Leben kaum wegzudenken – umso wichtiger ihre Sicherheit! Auf der AI.BAY 2023 am 24./25.02.2023 stellen wir unsere Forschungsergebnisse zur Manipulation & Absicherung von KI vor: #Deepfake #AdversarialExamples Kostenlos zum Online-Event: aisec.fraunhofer.de/de/press…

Demystifying the Transferability of Adversarial Attacks in Computer Networks ieeexplore.ieee.org/abstract… #cybersecurity #MachineLearning #DeepLearning #AI #ArtificialIntelligence #IEEE #TNSM #IEEETNSM #AdversarialExamples #AdversarialAttacks

残念ながらAdversarialExamplesを学習データに混ぜるともっと強固なAIができてしまうのでむしろ逆効果なんですよね 学習時に問題を起こすのではなく推論時に問題を起こすのでは手法なので

But that's the *least* esoteric of the concerns about ML judgments. Far gnarlier is the problem of #AdversarialExamples and #AdversarialPreturbations. 4/

ディープラーニングベースのAVエンジンを回避するマルウェアを作成する手法。従来手法はAVエンジンに複数回のクエリを送る必要があったが、本手法は1回の送信で作成可能。本研究分野はAdversarial Malware example Generation (AMG)と呼ばれるとのこと。 #adversarialexamples arxiv.org/abs/2112.01724

AIセキュリティ情報発信ポータル更新のお知らせです。今回は、AIへの入力データに微小のノイズを加えることでAIの判断を誤らせる「敵対的サンプル」と、その防御手法である「敵対的学習」を解説しています。 #aisec #adversarialexamples mbsd.jp/aisec_portal/

入力画像のブラー部分（ボケた部分）に摂動を加えることで、画像分類器に誤分類を引き起こさせる攻撃手法。ブラー部分の摂動は目立たず、入力画像はより自然に見えるため、攻撃検知が困難になる。 #adversarialexamples arxiv.org/abs/2111.12971

Adversarial attacks on deep learning in computer vision - a comprehensive literature review @ResearchGate: researchgate.net/publication… #DeepLearning #AdversarialExamples #survey

📣 Start of the #ML2R #Blog series on robust #MachineLearning against Adversarial Attacks! First up, Linara Adilova details the workings of #AdversarialExamples, manipulated input signals designed to impair the classification function of #ML #algorithms.➡️ machinelearning-blog.de/grun…

Advertorch: A Toolbox for Adversarial Robustness Research #pytorch #adversarialexamples github.com/BorealisAI/advert…

Deepfake detectors can be defeated, computer scientists show for the first time techxplore.com/news/2021-02-… via @techxplore_com #AI #MachineLearning #Deepfakes #AdversarialExamples @UCSanDiego

Hot off the digital press! The NBNN Daily is out! nbnn.info/news.html Thanks to @RichardGarriott #adversarialexamples #data