CIFAR-10
PulseAugur coverage of CIFAR-10 — every cluster mentioning CIFAR-10 across labs, papers, and developer communities, ranked by signal.
- instance of CIFAR-100 70%
- instance of Tiny-ImageNet 70%
- used by federated learning 70%
- instance of residual neural network 70%
- instance of Fashion-MNIST 70%
- used by SGD 70%
- used by residual neural network 70%
- instance of ImageNet ILSVRC-2012 70%
- instance of ImageNet-100 70%
- competes with AdamW 70%
- used by Imagenette 70%
- instance of differential privacy 70%
19 day(s) with sentiment data
-
Magnetic neuron enables signed spiking for richer AI data processing
Researchers have developed a new type of neuron using a magnetic tunnel junction (MTJ) that can process signed information, offering richer data representation than standard spiking neurons. This MTJ-based neuron mimics…
-
New research tackles adversarial robustness in deep neural networks
Several recent research papers explore novel methods for enhancing the adversarial robustness of deep neural networks. These studies introduce techniques such as ensemble-based approaches combining empirical and certifi…
-
FACT framework improves active finetuning for pretrained models
Researchers have introduced FACT, a novel framework designed to enhance the efficiency and effectiveness of active finetuning for pretrained models. This approach addresses the issue of feature distortion during finetun…
-
New $\ell_p$-norm scheme enhances deep learning optimization
Researchers have introduced a new optimization scheme for deep neural networks that utilizes a dynamic $\ell_p$-norm, moving beyond the limitations of fixed $\ell_2$ and $\ell_\infty$ norms. This novel approach, termed …
-
New NPPR metric offers robust deep learning evaluation
Researchers have introduced Non-Parametric Probabilistic Robustness (NPPR), a new metric for evaluating the robustness of deep learning models. Unlike previous methods that assume a known perturbation distribution, NPPR…
-
Graph curvature method enhances neural network pruning
Researchers have introduced a novel approach to neural network pruning by leveraging graph theory, specifically Ollivier-Ricci curvature (ORC). This method identifies critical data flows and connections within a neural …
-
Federated learning policy cuts IIoT training time and energy use
Researchers have developed a new bandwidth allocation policy for federated learning systems operating over industrial IoT networks. This policy partitions participating devices into ordered subsets, granting each subset…
-
New TASER framework boosts deep learning model robustness
Researchers have developed TASER, a new training framework called Task-Aware Stein Regularisation, designed to improve the robustness of deep learning models against distribution shifts and adversarial attacks. This met…
-
New framework unifies credit assignment for neural networks
Researchers have developed a new framework called Score Broadcast and Decorrelation (SBD) for credit assignment in neural networks. This framework is designed to work with various differentiable loss functions, offering…
-
Student capacity and architecture correctness key to knowledge distillation
A new study published on arXiv investigates the effectiveness of knowledge distillation (KD) in ResNet models for image classification on CIFAR-10. The research found that a student model's capacity significantly impact…
-
Federated Learning Research Tackles Privacy, Forgetting, and Heterogeneity
Recent research in federated learning (FL) addresses critical challenges in privacy and data drift. One paper introduces TADI and Fulcrum to protect against topology-aware inference attacks by optimally allocating noise…
-
New AHGC method improves out-of-distribution detection in AI
Researchers have developed a new method called Adaptive Hierarchical Graph Cut (AHGC) for out-of-distribution (OOD) detection in machine learning. This approach addresses the challenge of distinguishing between in-distr…
-
New HCL-FF framework boosts Forward-Forward algorithm for neural networks
Researchers have developed a new framework called HCL-FF to improve the Forward-Forward (FF) algorithm, a biologically plausible alternative to backpropagation for training neural networks. This enhanced method incorpor…
-
DASH framework generates stealthy adversarial AI examples
Researchers have developed DASH, a meta-attack framework designed to create adversarial examples for AI models that are both effective at causing misclassification and visually imperceptible. This framework strategicall…
-
New hierarchical method efficiently prunes CNN filters
Researchers have developed a novel two-level hierarchical approach for whole-network filter pruning in Convolutional Neural Networks (CNNs). This method efficiently reduces model size and computational requirements by p…
-
TSFLora framework cuts AI model adaptation costs for edge devices
Researchers have developed TSFLora, a novel framework designed to efficiently adapt large AI models for use on wireless edge devices. This method addresses the limitations of existing approaches like federated fine-tuni…
-
New diagnostic tool optimizes neural network pruning at high sparsity
Researchers have developed a new diagnostic tool called Relative Repairability (RR) to help optimize neural network pruning, particularly at high sparsity levels. RR assesses how much damage from pruning can be recovere…
-
New certificate method analyzes VAE constant collapse
Researchers have developed a new method to certify and analyze constant collapse in variational autoencoders (VAEs). This technique uses a simplex witness certificate to determine if the encoder mean becomes independent…
-
New framework uses approximate latent structure for certifiable classifier robustness
Researchers have developed a new framework to create certifiably robust deep learning classifiers by leveraging the latent structure within data representations. Their method proves that even approximate Gaussian mixtur…
-
New adaptive optimizer PILOT improves deep learning accuracy
Researchers have developed PILOT, a novel adaptive optimizer for deep learning that adjusts its update strategy during training. Unlike traditional optimizers with fixed update rules, PILOT uses gradient-direction agree…