DRIFT uses resilience analysis, targeted DVFS, and adaptive rollback ABFT to deliver 36% average energy savings or 1.7x speedup in diffusion model inference while preserving generation quality.
Approxabft: Approximate algorithm-based fault tolerance for vision transformers
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
verdicts
UNVERDICTED 2representative citing papers
An adaptive vulnerability-aware fault tolerance framework for neural networks that employs a GNN predictor to dynamically adjust protection policies, achieving over 95% prediction accuracy and 42.12% average overhead reduction.
citing papers explorer
-
DRIFT: Harnessing Inherent Fault Tolerance for Efficient and Reliable Diffusion Model Inference
DRIFT uses resilience analysis, targeted DVFS, and adaptive rollback ABFT to deliver 36% average energy savings or 1.7x speedup in diffusion model inference while preserving generation quality.
-
Adaptive Soft Error Protection for Neural Network Processing
An adaptive vulnerability-aware fault tolerance framework for neural networks that employs a GNN predictor to dynamically adjust protection policies, achieving over 95% prediction accuracy and 42.12% average overhead reduction.