Self-Supervised Neighborhood Probing Shields Tabular AI From Poisoned Labels
Researchers at Soongsil University have developed a self-supervised, multi-view defense that detects label-flipping poisoning attacks in tabular machine learning by ...
Tag: data poisoning
Machine Learning’s Blind Spot: The Hidden Adversarial Threats to Unsupervised AI
A systematic review of 93 studies reveals eight attack types and nine defence types across six categories of unsupervised machine ...
Four Ways Hackers Can Break Into Your Voice: The New Science of Audio Fraud
A new survey in Artificial Intelligence Review unifies the four major attack vectors against voice authentication, finding that audio deepfakes ...
Hackers Can Hijack Graph AI With Just a Handful of Poisoned Samples
Researchers have unveiled a data-level prompt injection attack that hijacks graph prompt learning systems by injecting a tiny fraction of ...
Flat Gradients Make Fake Users Deadlier: Smarter Attacks Expose Recommender Vulnerabilities
Researchers have developed RecGP, a gradient penalty method that crafts transferable adversarial fake users by steering them toward flat regions ...
-
Mothers who receive childcare support from maternal grandparents show more parental warmth, finds NTU Singapore study
-
University of Seville Breaks 120-Year-Old Mystery, Revises a Key Einstein Concept
-
Bee body mass, pathogens and local climate influence heat tolerance
-
Researchers record first-ever images and data of a shark experiencing a boat strike
-
Groundbreaking Clinical Trial Reveals Lubiprostone Enhances Kidney Function
Embark on a thrilling journey of discovery with Scienmag.com—your ultimate source for cutting-edge breakthroughs. Immerse yourself in a world where curiosity knows no limits and tomorrow’s possibilities become today’s reality!