Latest Articles for Deep Learning

A look into analog deep learning's potential benefits and challenges in artificial intelligence.

This article discusses new methods improving deep learning performance using nonlocal derivatives.

A new approach improves realism and variety in generated images using deep learning.

A new tool enhances segmentation accuracy for 7T MRI brain images.

FastSAM-3DSlicer simplifies 3D segmentation in medical images for improved diagnosis.

CoSIGN offers rapid and effective solutions for image restoration challenges.

Examining the vulnerability of neural networks to small input changes.

A two-stage approach improves the accuracy of brain metastases segmentation using MRI.

A new method to select data augmentations improves model performance on time series tasks.

Examining adversarial attacks and model robustness in semantic segmentation.

A new method enhances the security of deep learning models against hidden threats.

A new method enhances tumor detection accuracy using weakly supervised learning techniques.

New methods improve accuracy in medical image segmentation across varied data sources.

ConvBench offers a new way to assess convolution algorithm performance effectively.

This research aims to clarify the mechanics of GCP and Riemannian metrics.

New methods improve genetic operators in evolutionary algorithms using deep learning.

Introducing a new method for improving light field image resolution.

Tiled Bit Networks optimize deep learning models for better efficiency and performance.

Relational Representation Distillation improves model efficiency and accuracy in knowledge transfer.

A new weight generator improves neural network training efficiency and quality.

Examining the assumptions in deep learning optimization methods and their real-world implications.

Introducing NAMER, a new method for recognizing handwritten math expressions with improved speed and accuracy.

MCU-MixQ enhances AI model performance on microcontrollers by optimizing resource use.

This article examines how embedding initialization affects transformer model performance.

A novel pooling method enhances AI for better cancer detection.

A new method to represent attention models through clear diagrams to enhance understanding.

A new method improves uncertainty estimation in machine learning applications.

Explore the benefits and dynamics of using Poisson SGD for model training.

A new model combining Unet and TransUnet for improved nuclei segmentation.

Study shows deep learning's role in improving aortic segmentation in CT images.

Introducing SPLITZ, a method for improving AI model stability against adversarial examples.

A new framework improves deep learning training by integrating hardware and task management.

MetaAug reduces overfitting in PTQ through innovative data transformations.

SparseCraft reshapes 3D modeling using fewer images for better results.

Exploring how DNNs recover functions in overparameterized settings.

DeepLabv3++ model improves accuracy in colorectal cancer screening.

New method enhances deep learning models for limited-resource devices.

New methods improve image recovery from blurriness with deep learning techniques.

GSN combines dictionary learning and graph techniques for improved medical image classification.

A new method enhances object detection in remote sensing images.