batch normalization
PulseAugur coverage of batch normalization — every cluster mentioning batch normalization across labs, papers, and developer communities, ranked by signal.
1 day(s) with sentiment data
-
Machine unlearning evaluations flawed by BatchNorm artifact
Researchers have identified a significant artifact in machine unlearning evaluations, particularly affecting models that use Batch Normalization (BatchNorm). This artifact, termed the "BatchNorm Illusion," can reverse a…
-
New FORGE method enables test-time adaptation for integer-only vision models on microcontrollers
Researchers have developed FORGE, a novel forward-only test-time adaptation method specifically designed for integer-only vision models running on microcontrollers. This method addresses the challenge of adapting models…
-
New gradient-free method enables test-time adaptation for frozen AI models
Researchers have developed CASTER, a novel gradient-free method for test-time adaptation (TTA) that allows models to adapt without updating their parameters. This approach is particularly useful for inference-only accel…
-
New technique stabilizes Proximal Policy Optimization training
Researchers have developed a new technique called Mode-Dependent Rectification (MDR) to stabilize Proximal Policy Optimization (PPO) training in reinforcement learning. This method addresses issues caused by mode-depend…
-
New research explores finite-difference methods for PINNs
A new paper explores the use of finite-difference (FD) methods for computing derivatives in Physics-Informed Neural Networks (PINNs), presenting it as an alternative to automatic differentiation (AD). The research demon…
-
Finite-difference methods offer faster, more accurate derivatives for PINNs
A new paper explores finite-difference (FD) methods as an alternative to automatic differentiation (AD) for computing derivatives in physics-informed neural networks (PINNs). The research indicates that FD can match AD …
-
New SNN training and pruning methods boost efficiency and performance
Researchers are developing new methods to improve the efficiency and performance of Spiking Neural Networks (SNNs). One approach, Criticality-Constrained Quadratic Pruning (CQP), uses a combination of weight magnitude a…
-
SaluNet replaces normalization layers with learnable activation
Researchers have developed SaluNet, a novel deep network architecture that eliminates the need for traditional normalization layers like BatchNorm and LayerNorm. This is achieved through a new learnable activation funct…
-
Batch Normalization increases AI model memorization and privacy risks
A new research paper published on arXiv explores how Batch Normalization (BN) in deep neural networks can inadvertently increase the risk of data memorization and privacy breaches. The study found that BN significantly …
-
StableGrad stabilizes deep neural network training without batch normalization
Researchers have introduced StableGrad, a novel optimizer-level mechanism designed to control the scale of activations and gradients in deep neural networks. This method aims to prevent training instability without rely…
-
New research explores batch normalization's geometric impact on neural network partitions
Two new research papers explore advancements in Batch Normalization (BN) for neural networks. One paper investigates how training-time BN affects the geometric partitioning of functions in piecewise-affine networks, sug…