convolutional neural network
PulseAugur coverage of convolutional neural network — every cluster mentioning convolutional neural network across labs, papers, and developer communities, ranked by signal.
6 天有情绪数据
-
New SISA framework enables efficient class-level machine unlearning in CNNs
Researchers have developed a novel machine unlearning technique specifically for removing entire classes of data from deep neural networks. This method modifies the Sharded, Isolated, Sliced, and Aggregated (SISA) frame…
-
GourNet CNN model achieves 97% accuracy in mango leaf disease detection
Researchers have developed GourNet, a Convolutional Neural Network model designed to detect diseases in mango leaves. Trained on the MangoLeafBD dataset, which includes eight classes (seven diseases and one healthy), Go…
-
New MANN method enhances gradient boosting with neural networks for diverse data
Researchers have introduced Multiple Additive Neural Networks (MANN), a novel methodology that replaces decision trees with shallow neural networks in the Gradient Boosting framework. This approach integrates Convolutio…
-
AI study finds lung segmentation vital for COVID-19 X-ray diagnosis
A new study published on arXiv investigates the necessity of data augmentation and lung segmentation for AI-driven COVID-19 detection using chest X-rays. The research, which proposes a methodology called SDL-COVID, foun…
-
迈向可解释的人工智能：利用量子退火进行特征选择
研究人员开发了一种新颖的方法，通过利用量子退火进行特征选择来解释卷积神经网络（CNN）在图像分类任务中的应用。该方法识别出对模型预测最有影响的特征图，旨在提高AI系统的透明度和可信度。该技术将特征选择问题编码为量子约束优化问题，然后使用量子退火进行求解。评估结果显示，与现有的可解释AI方法（如GradCAM和GradCAM++）相比，类解纠缠得到了改善。
-
New OCR pipeline enhances retail bill digitization with adaptive enhancement
Researchers have developed and benchmarked an adaptive Optical Character Recognition (OCR) pipeline specifically designed for digitizing diverse retail bills. This system incorporates a CNN-based enhancement module, an …
-
新框架通过分离层优化深度学习训练
研究人员引入了一个名为层分离优化（Layer Separation Optimization）的新颖框架，以解决深度学习模型使用交叉熵损失进行训练时面临的挑战。该方法旨在缓解深度网络训练过程中出现的强非凸性问题。通过使用辅助变量将复杂的优化问题分解为更小、更易于管理子问题，该框架在理论上为原始交叉熵损失提供了上限，并在数值实验中展示了改进的优化行为。
-
AI identifies writers of historical Arabic manuscripts with high accuracy
Researchers have developed a Convolutional Neural Network (CNN) with attention mechanisms to identify writers of historical Arabic manuscripts. The study, using the Muharaf dataset, expanded writer labels and establishe…
-
CNNs and personalized thresholds improve driver drowsiness detection accuracy
Researchers have developed a new driver drowsiness detection system that uses personalized thresholds for Eye Aspect Ratio (EAR) and Mouth Aspect Ratio (MAR) to account for individual differences. The system integrates …
-
Machine learning models reveal geographic data improves insurance claim predictions
Researchers have developed a method to incorporate geographic information into motor insurance claim prediction models, even with limited location data. By utilizing environmental data from OpenStreetMap and CORINE Land…