GPU memory accesses on Apple M1 leave observable footprints in the shared system-level cache, letting an unprivileged CPU process recover sensitive GPU workload data.
Badnl: Backdoor attacks against NLP models with semantic-preserving improvements,
6 Pith papers cite this work, alongside 82 external citations. Polarity classification is still indexing.
6
Pith papers citing it
82
external citations · external index
citation-role summary
background 1
citation-polarity summary
roles
background 1polarities
background 1representative citing papers
A systematic study of CPU-GPU memory sharing on four UMA platforms yields four performance findings and a library that automatically selects correct, efficient implementations.
CapsLock uses a revoke-on-use hardware capability mechanism to detect Rust principle violations at machine-code level, covering FFI and inline assembly that Miri cannot handle.
BadSem shows that semantic mismatches between images and text can serve as stealthy backdoor triggers for VLMs, achieving near-perfect attack success with low poisoning rates.
MorphStrata generates heterogeneous student models via layer-specific perturbations in a Transformer-based Morphence MTD setup, reporting RMSE gains up to 24% and 98% on AEP data under FGSM and BIM attacks with under 1% training time increase.
A random forest trained on a new inter-flow signal representation reports high accuracy on eight encrypted-traffic datasets, but test-set hyperparameter selection and likely chunk leakage undermine the claim.
citing papers explorer
-
SLAC: Access-Driven CPU-to-GPU Side-channel Attacks via System-Level Cache on Apple Silicon
GPU memory accesses on Apple M1 leave observable footprints in the shared system-level cache, letting an unprivileged CPU process recover sensitive GPU workload data.
-
Characterizing CPU-GPU Memory Sharing over Unified Memory Architecture
A systematic study of CPU-GPU memory sharing on four UMA platforms yields four performance findings and a library that automatically selects correct, efficient implementations.
-
Securing Mixed Rust with Hardware Capabilities
CapsLock uses a revoke-on-use hardware capability mechanism to detect Rust principle violations at machine-code level, covering FFI and inline assembly that Miri cannot handle.
-
Backdoor Attack on Vision Language Models with Stealthy Semantic Manipulation
BadSem shows that semantic mismatches between images and text can serve as stealthy backdoor triggers for VLMs, achieving near-perfect attack success with low poisoning rates.
-
MorphStrata: Layer-Specific Perturbations for Generating Morphence Students in Time-Series Moving Target Defense
MorphStrata generates heterogeneous student models via layer-specific perturbations in a Transformer-based Morphence MTD setup, reporting RMSE gains up to 24% and 98% on AEP data under FGSM and BIM attacks with under 1% training time increase.
-
Generalized Encrypted Traffic Classification Using Inter-Flow Signals
A random forest trained on a new inter-flow signal representation reports high accuracy on eight encrypted-traffic datasets, but test-set hyperparameter selection and likely chunk leakage undermine the claim.