ResNet-18
PulseAugur coverage of ResNet-18 — every cluster mentioning ResNet-18 across labs, papers, and developer communities, ranked by signal.
5 day(s) with sentiment data
-
UAV audio classification: Method scaling beats model scaling
A new research paper explores the trade-offs between model size and fine-tuning methods for audio classification on unmanned aerial vehicles (UAVs). The study found that parameter-efficient fine-tuning (PEFT) methods, p…
-
FPGA platform accelerates approximate multiplier evaluation for DNNs
Researchers have developed FAME, a new platform utilizing FPGAs to accelerate the evaluation of approximate multipliers for deep neural networks. This hardware-based approach significantly reduces the time needed to ass…
-
New watermarking scheme TwinMark protects AI models from distillation attacks
Researchers have developed TwinMark, a novel watermarking technique designed to protect AI models against distillation attacks. This method uses two complementary linear functionals, one based on feature covariance and …
-
New method enhances deep neural network interpolation robustness
Researchers have introduced Sharp Mode Connectivity (SMC), a new method for optimizing parametric curves in the weight space of deep neural networks. Unlike standard mode connectivity, which only ensures low loss along …
-
New REQAP method boosts DNN efficiency and resilience on edge devices
Researchers have developed REQAP, a novel methodology for optimizing Deep Neural Networks (DNNs) on edge accelerators. This approach combines a reliability-aware mixed-precision quantization framework with a determinist…
-
New research compares OCR-based VQA systems against end-to-end models under image degradation
A new research paper explores the effectiveness of text-centric Visual Question Answering (VQA) systems when images are degraded by common issues like blur or low resolution. The study compares modular OCR-based pipelin…
-
New research explores AI model architectures and tokenization for ECG analysis
Two new research papers explore architectural and tokenization strategies for improving AI models in analyzing electrocardiograms (ECGs). The first paper introduces R-U-Net, which enhances ECG delineation by optimizing …
-
New tunable lifting schemes improve ResNet-18 performance in image tasks
Researchers have developed a new family of tunable lifting schemes for biorthogonal wavelet filter banks, offering three distinct strategies for adapting low-pass, high-pass, or both frequency branches. These schemes ar…
-
New benchmark Infra-Bench CLS tests foundation models on critical infrastructure classification
A new benchmark, Infra-Bench CLS, has been introduced to evaluate the effectiveness of Earth observation foundation models in classifying critical infrastructure. The benchmark comprises 18,756 images of facility-scale …
-
FANS framework optimizes model architectures for heterogeneous federated learning
Researchers have developed FANS (Federated Adaptive Network Search), a new framework designed to optimize model architectures in heterogeneous federated learning environments. This approach utilizes a hypernetwork to le…
-
Phase transition frequency predicts ResNet accuracy in training
Researchers have identified a new metric, "phase transition frequency," that can predict the test accuracy of ResNet models during training. This metric, which counts discrete class-separability jumps, showed a strong n…
-
New ARC-CT framework enhances 3D chest CT analysis with vision-language learning
Researchers have developed ARC-CT, a novel framework for contrastive vision-language learning specifically designed for 3D chest CT scans and radiology reports. This approach addresses limitations in standard contrastiv…
-
New research explores advanced machine unlearning techniques for AI models · 4 sources tracked
Researchers are developing new methods for machine unlearning, the process of removing specific data or knowledge from AI models. One approach, Source-Free Class Relearning Audit (SFRA), focuses on recovering forgotten …
-
New AI model AURASeg enhances drivable area segmentation for robots
Researchers have developed AURASeg, a novel segmentation framework designed to improve the accuracy of identifying drivable areas for autonomous robots. This framework addresses limitations in conventional models by enh…
-
GradAttn enhances CNNs with attention-modulated gradient flow
Researchers have introduced GradAttn, a novel approach to enhance deep convolutional neural networks (CNNs) by replacing fixed residual connections with attention-controlled pathways. This method dynamically weights fea…
-
New DeltaMomentum optimizer speeds up deep learning training
Researchers have introduced DeltaMomentum, a novel approach to updating momentum in deep learning optimizers. Unlike traditional methods that use a fixed rate for exponential moving averages, DeltaMomentum dynamically a…
-
New technique boosts differentially private training accuracy for vision models
Researchers have developed a new technique called Spectral Gradient Orthogonalization (SGO) to improve the accuracy of differentially private training for vision models. This method addresses the issue where isotropic G…
-
Cross-validation improves hyperparameter tuning for medical image AI
A new research paper explores hyperparameter optimization (HPO) for deep learning image classifiers, particularly in medical imaging where small datasets are common. The study compared three HPO protocols: fixed holdout…
-
New research compares training objectives for AI out-of-distribution detection
A new research paper systematically compares four training objectives for out-of-distribution (OOD) detection in image classification. The study evaluated Cross-Entropy Loss, Prototype Loss, Triplet Loss, and Average Pr…
-
New Uniform Herding method improves class-incremental learning
Researchers have introduced Uniform Herding, a novel method for exemplar replay in class-incremental learning. This technique aims to preserve performance on earlier classes as feature representations evolve by managing…