CIFAR-100
PulseAugur coverage of CIFAR-100 — every cluster mentioning CIFAR-100 across labs, papers, and developer communities, ranked by signal.
8 天有情绪数据
-
New MDSE attack fools Spiking Neural Networks and traditional models
Researchers have developed a new adversarial attack method called Mixed Dynamic Spiking Estimation (MDSE) specifically for Spiking Neural Networks (SNNs). This attack demonstrates that the effectiveness of white-box adv…
-
Research explores how sparsity allocation affects neural network recovery after pruning
A new research paper investigates how the allocation of sparsity in neural networks impacts their ability to recover accuracy after pruning, especially when labeled retraining data is unavailable. The study compares dif…
-
LLM multi-agent system automates neural network customization for MCUs
Researchers have developed AutoMCU, a novel system that leverages LLM-based multi-agent approaches to customize neural networks for microcontroller units (MCUs). This method prioritizes feasibility by integrating vendor…
-
FAIR-Pruner framework enables adaptive layer-wise neural network pruning
Researchers have developed FAIR-Pruner, a new framework designed for automatic, layer-wise structured pruning of deep neural networks. This method adaptively allocates sparsity across network layers by using both remova…
-
SpikingMoE integrates Mixture-of-Experts into spike-driven Transformers
Researchers have introduced SpikingMoE, a novel framework that combines Spiking Neural Networks (SNNs) with a Mixture-of-Experts (MoE) architecture. This approach utilizes a spike-driven prompt (SDprompt) for biological…
-
New EIHF method boosts OOD detection in vision models
Researchers have developed a new method called Early High-Frequency Injection (EIHF) to improve out-of-distribution (OOD) detection in computer vision models. EIHF works by injecting high-frequency information into the …
-
HamJEPA advances JEPAs with Hamiltonian geometry and symplectic prediction
Researchers have introduced HamJEPA, a novel approach to Joint Embedding Predictive Architectures (JEPAs) that moves beyond isotropic regularization. This new method encodes views as phase-space states and uses a learne…
-
CutMix training protocol induces spatial locality in Vision Transformers
Researchers have found that specific training techniques can encourage spatial locality in Vision Transformers. By using a 'Modern' protocol involving data augmentation like CutMix and ColorJitter, along with label smoo…
-
Federated Martingale Posterior sampling improves Bayesian neural networks
Researchers have introduced Federated Martingale Posterior (FMP) sampling, a novel protocol for federated Bayesian neural networks. This method addresses the difficulty of specifying priors in large models by using a pr…
-
New conformal prediction methods improve uncertainty quantification
Two new research papers introduce novel approaches to conformal prediction, a method for quantifying uncertainty in machine learning models. The first paper, "Decoupled Conformal Optimisation," proposes a train-tune-cal…
-
Neural networks learn image features via Fourier analysis
Researchers have explored the learning dynamics of neural networks through a Fourier perspective, focusing on how they learn simpler features before more complex ones. Their work introduces a synthetic data model for tr…
-
New MIND framework tackles model-induced label noise
Researchers have introduced MIND, a novel framework designed to tackle model-induced label noise in machine learning. This noise arises from the inherent biases of pre-trained models used for data annotation, leading to…
-
Code embeddings boost neural architecture search efficiency
Researchers have developed a novel method called Code-Oriented LM Embeddings (COLE) to improve Neural Architecture Search (NAS). This technique uses off-the-shelf language models to generate embeddings from code represe…
-
New TILT method improves unsupervised domain adaptation
Researchers have introduced Target-Induced Loss Tilting (TILT), a new method for unsupervised domain adaptation that addresses covariate shift. TILT utilizes a novel objective function to train a source predictor by pen…
-
New OUIDecay method adapts CNN regularization layer-by-layer
Researchers have introduced OUIDecay, a novel adaptive weight decay method for convolutional neural networks. This technique dynamically adjusts regularization strength for each layer based on online activation patterns…
-
New certificate method detects constant collapse in VAEs
Researchers have developed a new method to detect and prevent a specific type of failure in variational autoencoders (VAEs) known as constant collapse. This technique provides a testable certificate that can distinguish…
-
New AS-LoRA method improves privacy in federated learning
Researchers have developed AS-LoRA, a novel framework for adaptive selection of LoRA components in privacy-preserving federated learning. This method addresses aggregation errors common in such setups by allowing each l…
-
New parameter E predicts Mixture-of-Experts model health, preventing dead experts.
Researchers have introduced a new dimensionless control parameter, E = T*H/(O+B), to predict the health of expert ecologies in Mixture-of-Experts (MoE) models. This parameter, derived from four hyperparameters, can prev…
-
Hierarchy-Aware Cross-Entropy improves image classification accuracy
Researchers have introduced Hierarchy-Aware Cross-Entropy (HACE), a novel loss function designed to improve image classification by accounting for semantic relationships between classes. Unlike standard cross-entropy, H…
-
AI research tackles layer free-riding and enhances data privacy for models
Researchers have identified a phenomenon in Forward-Forward networks called layer free-riding, where later layers can inherit tasks already partially handled by earlier layers, leading to a decay in gradient. Three loca…