Data sensitivity shaped Chinese adults’ health data-sharing choices
A stated-choice experiment with 1,481 Chinese adults found that data sensitivity strongly changed which hypothetical health data-sharing arrangements participants selected, ...
Tag: privacy
Hybrid encryption protects federated updates in Medical-MNIST test
A Paillier-CKKS framework encrypted federated-learning updates and reached 94.67% average accuracy on Medical-MNIST during simulated reverse-engineering attacks.
Your Phone Knows When You Are Stressed: KAIST Turns Everyday App Habits Into a Mental Health Signal
KAIST researchers developed a technology called SHIRBT-MMF that identifies daily stress from repeated smartphone behaviors such as messenger chatting, web ...
Federated learning brings privacy-first anomaly detection to smart grids
Smart grids promise a cleaner, more responsive electricity system, but they also carry a hidden vulnerability: every smart meter is ...
AI Learns to Offload Your Data Without Spilling Your Secrets
A deep reinforcement learning method called P2TO extends task offloading from edge and cloud servers down to individual virtual machines, ...
Geometry Meets the Spreadsheet: Flow Matching Reshapes Synthetic Tabular Data
A new flow matching framework assigns each tabular variable type its natural geometric space, improving synthetic data quality and resisting ...
The Drafting Fiction: How Regulators Really Govern AI Scribes in Healthcare
A new analysis of 74 official documents finds that AI scribe governance hinges on a fragile convention that machine-generated notes ...
AI Has Conquered Saudi Arabia’s Media Industry, But Not Equally
A survey of 400 Saudi audio-visual media professionals finds near-saturated AI adoption, with privacy, cost, and Arabic-language processing emerging as ...
Teaching Graph AI to Forget: SURGE Rewrites the Rules of Machine Unlearning
Researchers have developed SURGE, a graph unlearning framework that treats data deletion requests as structural perturbations and repairs graph neural ...
When Language Models Meet Graph Networks: New Map Charts the Trust Fault Lines
A systematic review in Applied Intelligence introduces a two-axis taxonomy that maps how large language models enhance graph neural networks ...
CoughNet: Low-cost sensors count coughs in a room while keeping identities private
Binghamton University researchers have built CoughNet, a low-cost Raspberry Pi and LoRa sensor network that counts unique coughers in a ...
Deep Learning in Medicine Hits a Wall: Explainability and Privacy Are Locked in Tension
A new review of 61 studies reveals that the explainability and privacy protections required for trustworthy medical and biometric deep ...
Digital Privacy: Health Data Control in Incarceration
In a groundbreaking study that is likely to reshape discussions around health information access, researchers have expressed deep concerns regarding ...
POPULAR NEWS
-
Intranasal Lidocaine Spray Eases Nasogastric Tube Placement in Adult Emergency Patients
-
Hospital Outbreaks Disseminate Ceftazidime-Avibactam and Cefiderocol Co-Resistant Klebsiella Pneumoniae Clones
-
PET Imaging Shows Aging Reduces Human Brain Ascorbic Acid Transport via Choroid Plexus
-
Plant Hemoglobin Expression Impairs Aerobic Guaiacol Bioconversion in Pseudomonas putida