Results 241 to 250 of about 183,434 (340)

Performance Analysis and Comparison between the Uncoded OIDMA and Convolutional Coded OIDMA

, 2014
Iosr Journals, Dr.Anil Kumar Tushar Srivastava   +1 more
openalex   +1 more source

A performance analysis of convolutional autoencoder modified WaveGAN architectures for realistic 12 lead electrocardiogram synthesis. [PDF]

Sci Rep
Liu J, Chui KT, Lee LK, Lo K, Yang A, Fong EKS.   +5 more
europepmc   +1 more source

QS4D: Quantization‐Aware Training for Efficient Hardware Deployment of Structured State‐Space Sequential Models

Advanced Intelligent Systems, EarlyView.
Quantization‐aware training creates resource‐efficient structured state space sequential S4(D) models for ultra‐long sequence processing in edge AI hardware. Including quantization during training leads to efficiency gains compared to pure post‐training quantization.
Sebastian Siegel, Ming‐Jay Yang, Younes Bouhadjar, Maxime Fabre, Emre Neftci, John Paul Strachan   +5 more
wiley   +1 more source

Software for dataset-wide XAI: From local explanations to global insights with Zennit, CoRelAy, and ViRelAy. [PDF]

PLoS One
Anders CJ, Neumann D, Samek W, Müller KR, Lapuschkin S.   +4 more
europepmc   +1 more source

Enhancing Atomic‐Resolution in Electron Microscopy: A Deep Learning Denoiser Operating in the Frequency Domain

Advanced Intelligent Systems, EarlyView.
The newly developed AI‐automated Fast Fourier Transform denoising algorithm surpasses conventional real‐space methods by revealing even light atoms otherwise hidden in noisy backgrounds. Atomic resolution electron microscopy has become an essential tool for many scientific fields, when direct visualization of atomic arrangements and defects is needed ...
Ivan Pinto‐Huguet, Marc Botifoll, Xuli Chen, Martin Børstad Eriksen, Jing Yu, Giovanni Isella, Andreu Cabot, Gonzalo Merino, Jordi Arbiol   +8 more
wiley   +1 more source

SVTR-MG: an optical character recognition network for food packaging spray codes. [PDF]

Sci Rep
Pan S, Wang P.
europepmc   +1 more source

Upsampling DINOv2 Features for Unsupervised Vision Tasks and Weakly Supervised Materials Segmentation

Advanced Intelligent Systems, EarlyView.
Feature from recent image foundation models (DINOv2) are useful for vision tasks (segmentation, object localization) with little or no human input. Once upsampled, they can be used for weakly supervised micrograph segmentation, achieving strong results when compared to classical features (blurs, edge detection) across a range of material systems.
Ronan Docherty, Antonis Vamvakeros, Samuel J. Cooper   +2 more
wiley   +1 more source

Lightweight bearing fault diagnosis via decoupled distillation and low rank adaptation. [PDF]

Sci Rep
Petrosian O, Li P, He Y, Liu J, Sun Z, Fu G, Meng L.   +6 more
europepmc   +1 more source

Building an Intelligent Cardiovascular System Platform: Embedding Artificial Intelligence across All Facets of Cardiovascular Medicine

Advanced Intelligent Systems, EarlyView.
This paper presents an integrated AI‐driven cardiovascular platform unifying multimodal data, predictive analytics, and real‐time monitoring. It demonstrates how artificial intelligence—from deep learning to federated learning—enables early diagnosis, precision treatment, and personalized rehabilitation across the full disease lifecycle, promoting a ...
Mowei Kong, Xingwei Zhao, Qiuting Li, Yang Yu, Chunxiang Zhang   +4 more
wiley   +1 more source

gReLU: a comprehensive framework for DNA sequence modeling and design. [PDF]

Nat Methods
Lal A, Gunsalus L, Nair S, Biancalani T, Eraslan G.   +4 more
europepmc   +1 more source