ResNet-50
PulseAugur coverage of ResNet-50 — every cluster mentioning ResNet-50 across labs, papers, and developer communities, ranked by signal.
14 day(s) with sentiment data
-
Deep learning framework tracks liquid sheet breakup dynamics
Researchers have developed a novel two-stage deep learning framework to analyze the complex dynamics of liquid sheet breakup into ligaments and droplets. The first stage utilizes a Faster R-CNN model with a ResNet-50 ba…
-
New RL method improves AI face anti-spoofing generalization
Researchers have developed a novel reinforcement learning approach called Grounding-Aware Judge Reward Reinforcement Learning (GA-JRRL) to improve the out-of-domain generalization of multimodal large language models (ML…
-
New method profiles Hessian mass in machine learning models
Researchers have developed a new method for evaluating block-diagonal curvature models in machine learning. This approach measures the squared off-block fraction of the parameter Hessian and analyzes the within-group cu…
-
Astronomical image classifiers' accuracy influenced by context, not just morphology
A new study published on arXiv evaluates the performance of small-data astronomical image classifiers, revealing that factors beyond object morphology significantly influence accuracy. Researchers developed a dataset to…
-
Self-supervised pose estimation model for construction machines unveiled
Researchers have developed a novel self-supervised pose estimation model for construction machinery, utilizing a three-stage approach that includes contrastive learning. This method significantly reduces the need for la…
-
New FDSML method enhances vision pre-training without real images
Researchers have introduced Formula-Driven Supervised Metric Learning (FDSML), a novel approach to pre-training vision models without real images. Unlike previous methods that treat pre-training as classification, FDSML…
-
New FDSML method enables vision pre-training without real images
Researchers have developed Formula-Driven Supervised Metric Learning (FDSML), a novel approach for vision pre-training that bypasses the need for real images. Instead of relying on discrete labels, FDSML derives supervi…
-
Knowledge distillation creates compact AI for diabetic retinopathy grading
Researchers have developed a knowledge distillation technique to create a compact model for grading diabetic retinopathy from retinal fundus images. This method successfully distilled knowledge from larger teacher model…
-
Vision Transformers show robust Alzheimer's MRI classification across datasets
A new study published on arXiv evaluates the robustness of Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs) for classifying Alzheimer's disease from MRI scans. Researchers compared five architectures,…
-
Adapted distilled image generators boost classifier accuracy
Researchers have developed a method to improve the utility of distilled text-to-image models for visual learning. By adapting these fast generators to reach an intermediate state of their slower, base models, the genera…
-
New research offers guidelines for AI-driven animal identification marks
A new research paper published on arXiv details effective strategies for designing back marks to aid in the automated identification of animals, particularly pigs. The study utilized a ResNet-50 neural network trained t…
-
New $L_p$ correlation measures offer improved generalization bounds for randomized learning
Researchers have introduced a new family of dependence measures called vector-valued $L_p$-norm correlation ($V_α$) to better understand the generalization capabilities of randomized learning algorithms. This framework …
-
New LEAN technique enhances neural network efficiency and error tolerance
Researchers have developed a new technique called LEAN (Loss-Engineered Modulation) to improve the efficiency and error tolerance of neural networks. LEAN works by altering the distribution of weight sensitivity in trai…
-
New LocaLeJEPA framework improves self-supervised learning with geometric priors
Researchers have introduced LocaLeJEPA, a new framework for self-supervised learning that addresses representation collapse issues in Joint-Embedding Predictive Architectures (JEPA). Unlike previous methods tied to Eucl…
-
Deep learning models identify airway collapse in sleep apnea patients using snoring sounds
Researchers have developed deep learning models, specifically the Audio Spectrogram Transformer (AST) and ResNet-50, to identify multi-level upper airway collapses in obstructive sleep apnea (OSA) patients using snoring…
-
FPGA-based FALCON system ensures latency guarantees for autonomous driving
Researchers have developed FALCON, a novel system for camera-based autonomous driving that prioritizes latency-certifiable occupancy networks. This system, implemented on an AMD Kria KR260 FPGA, ensures that latency gua…
-
Edge AI system DefoliationCAM assesses grapevine pest damage in real-time
Researchers have developed DefoliationCAM (D-CAM), an edge AI system designed to assess insect defoliation severity in grapevines in real-time. The system utilizes a YOLOX-Nano model for cordon localization and a lightw…
-
New defense method shields ML models from black-box query attacks
Researchers have developed a new defense mechanism called Gradient Confusion Defense (GCD) to protect machine learning models from black-box score-based query attacks. This method, which does not require retraining the …
-
New method certifies detection of adversarial AI attacks
Researchers have developed a new method for detecting adversarial attacks on deep learning classifiers, focusing on certifying that malicious inputs are flagged rather than ensuring model robustness. This approach lever…
-
New integer-only training method slashes neural network memory use
Researchers have developed a novel training algorithm for deep neural networks that exclusively uses integer arithmetic, eliminating the need for floating-point operations in any stage of the training process. This meth…