Robustness

Certified Robustness: Fundamentals and Challenges
Overview of classical and State-of-the-Art approaches to certify ML models’ robustness.
Jan 24, 2023 6:00 PM NTR LABS Webinars
Aleksandr Petiushko Александр Петюшко
Slides Video
Certified Robustness: Fundamentals and Challenges
Smoothed Embeddings for Certified Few-Shot Learning
Geometrical approach for certified few-shot learning via randomized smoothing of templates.
Mikhail Pautov, Olesya Kuznetsova, Nurislam Tursynbek, Aleksandr Petiushko Александр Петюшко, Ivan Oseledets
PDF Cite Poster Slides Source Document arXiv OpenReview Conf
Smoothed Embeddings for Certified Few-Shot Learning
Certified Robustness via Randomized Smoothing over Multiplicative Parameters of Input Transformations
A novel method of certified robustness using Rayleigh dostribution for smoothing when transformations consist not additive but multiplicative group.
Nikita Muravev, Aleksandr Petiushko Александр Петюшко
PDF Cite Video Source Document DOI arXiv
Certified Robustness via Randomized Smoothing over Multiplicative Parameters of Input Transformations
CC-Cert: A Probabilistic Approach to Certify General Robustness of Neural Networks
A theoretical approach to estimate the certified robustness tested on different image transformations.
Mikhail Pautov, Nurislam Tursynbek, Marina Munkhoeva, Nikita Muravev, Aleksandr Petiushko Александр Петюшко, Ivan Oseledets
PDF Cite Poster Source Document DOI arXiv
CC-Cert: A Probabilistic Approach to Certify General Robustness of Neural Networks
Geometry-Inspired Top-K Adversarial Perturbations
A simple geometrical approach for crafting top-k adversarial perturbations.
Nurislam Tursynbek, Aleksandr Petiushko Александр Петюшко, Ivan Oseledets
PDF Cite Video Source Document DOI arXiv
Geometry-Inspired Top-K Adversarial Perturbations
Certified Robustness in High Dimensions
The overview of certified robustness problems in high dimensions.
Apr 29, 2021 12:00 AM SkolTech
Aleksandr Petiushko Александр Петюшко
Slides
Certified Robustness in High Dimensions
CNN Adversarial Robustness in the Real World
A holistic overview of Adversarial Robustness, including digital and real world, mostly for the face detection and identification.
Apr 15, 2021 12:00 AM Lobachevsky University
Aleksandr Petiushko Александр Петюшко
Slides Talk Interview In Russian
CNN Adversarial Robustness in the Real World
CNN Robustness research: Application to face detectors and face ID systems
A short overview of Adversarial Robustness, in both digital and real world, explicitly for the face detection and identification.
Feb 4, 2021 12:00 AM OpentTalks.AI 2021
Aleksandr Petiushko Александр Петюшко
Slides Slides (In Russian)
CNN Robustness research: Application to face detectors and face ID systems
AdvHat: Real-World Adversarial Attack on ArcFace Face ID System
A practical approach to generate forehand printed patch to fool the best FaceID system - Arcface - in real world.
Stepan Komkov, Aleksandr Petiushko Александр Петюшко
PDF Cite Code Poster Slides Video Source Document DOI arXiv Demo
AdvHat: Real-World Adversarial Attack on ArcFace Face ID System
Robustness Threats of Differential Privacy
Surprising empirical observation showing how Differential Privacy hurts Adversarial Robustness.
Nurislam Tursynbek, Aleksandr Petiushko Александр Петюшко, Ivan Oseledets
PDF Cite arXiv Conf WorkshopWeb
Robustness Threats of Differential Privacy