A new way to find multiple system weaknesses at once
Researchers developed a two-layer method to estimate how many tests are needed to discover various safety violations.
Researchers developed a two-layer method to estimate how many tests are needed to discover various safety violations.
VellSSLH selectively adds security to code to prevent data leaks while keeping programs fast.
Researchers found that environmental changes and aging can allow attackers to bypass protective measures in FPGAs.
A system using a 4B-parameter model found 187 memory-safety findings across 4,000 projects.
Sable provides a way to prove that a specific model produced a given sequence of tokens bit for bit.
A new framework called SEAL modifies only 5% of facial image locations to disrupt identity acquisition.
A framework called DW-FedADMM uses dynamic weighting to help devices coordinate during threat defense.
Researchers studied how 19 defectors understand and adapt to digital privacy risks.
Current methods for integrating new security standards lack coordination across the software development lifecycle.
A belief-driven scheme uses a Bayesian security game to find hidden jamming UAVs.
The Zirah device uses noise to create random numbers for post-quantum security.
New randomized algorithms improve the speed and space needed for discrete Gaussian sampling and the Shortest Vector Problem.
Three formats show different performance levels when issuing, presenting, and verifying data.
A new mechanism uses dummy points to hide batch sizes and maintain privacy guarantees during data truncation.
A new cryptographic framework uses masked high-memory convolutional codes in the Niederreiter paradigm.
PairWatch uses satellite pairs and motion symmetry to find adversarial maneuvers.
The UniPCH framework separates payment logic from blockchain execution to enable cross-chain transactions.
Practical quantization levels have limited impact on membership privacy risks.
Host-based intrusion detectors order events and attack stages by timestamps but often leave assumptions about clocks and evidence unstated.
No tested mechanism fully compensates for the effect of large key sizes on performance.
The MIG-COW framework uses gradient decomposition to create adversarial examples for deepfake detectors.
Researchers developed OP-Print to identify if specific training examples were used in on-policy self-distillation.
Researchers analyzed recent progress in structural attacks on binary Goppa codes.
Researchers found that certain mathematical problems used in cryptography are less secure than previously estimated.