Results 251 to 260 of about 7,097,220 (297)

A Multi‐Structural Radiomic Digital Phenotyping Framework for Deciphering Preclinical Evolution of Degenerative Meniscal Tears

open access: yesAdvanced Science, EarlyView.
Subclinical matrix disruption precedes degenerative meniscal tears across the knee joint. A spatial compartmental radiomics model utilizing 3D‐DESS MRI captures early osteochondral destabilization, accurately predicting incident tears up to four years pre‐collapse. Feature attributions uncover coordinated chondral and osseous alterations, advancing non‐
Shengfa Li   +13 more
wiley   +1 more source

Device Stochasticity Directly Implements Bayesian Sampling in Spiking Neurons

open access: yesAdvanced Science, EarlyView.
Brain‐inspired stochastic spiking turns device noise into Bayesian uncertainty for robust, low‐power neuromorphic computing. ABSTRACT Device stochasticity is considered a non‐ideality in neuromorphic hardware, although stochastic neural activity supports probabilistic computation in biological systems.
Huannan Zheng, Jingli Liu, Kezhou Yang
wiley   +1 more source

ReLink: Reliable Deep Learning for Microscopic Motion Tracking

open access: yesAdvanced Science, EarlyView.
ReLink enables reliable trajectory linking in single‐particle tracking by integrating physics‐informed motion statistics with confidence‐based adaptive filtering. Across diverse simulated and experimental conditions, ReLink robustly reconstructs trajectories and preserves the fidelity of downstream analyses, enabling accurate characterization of ...
Yuan Jiang   +7 more
wiley   +1 more source

Enhanced High Dimensionality and the Information Processing Capacity in Interfered Spin Wave‐Based Reservoir Computing, Achieved With Eight Detectors

open access: yesAdvanced Electronic Materials, EarlyView.
Physical reservoir computing (PRC) based on spin wave interference has demonstrated high computational performance, yet room for improvement remains. In this study, we fabricated this concept PRC with eight detectors and evaluated the impact of the number of detectors using a chaotic time series prediction task.
Sota Hikasa   +6 more
wiley   +1 more source

AI‐Assisted Ultrasensitive Biosensing using Metal‐Electrolyte‐Metal‐Insulator‐Silicon Structure Based on Oxygen‐Tunable Iridium Oxide Nanonets for Lysyl‐Oxidase‐Like‐2 Breast Cancer Detection

open access: yesAdvanced Electronic Materials, EarlyView.
An oxygen‐controlled iridium‐oxide (IrOx) nano‐net structure is developed to enhance the sensitivity of metal‐electrolyte‐metal‐insulator‐silicon (MEMIS) biosensors. Integrated with a deep learning model, this platform achieves a high accuracy of 94.8% for detecting the breast cancer biomarker LOXL2.
Chiao‐Fan Chiu   +9 more
wiley   +1 more source

Predictable and Scalable Analog Matrix–Vector Multiplication in Memristor Crossbars via Closed‐Form Wire‐Resistance Compensation

open access: yesAdvanced Electronic Materials, EarlyView.
Wire resistance and sneak paths severely compromise the performance of large memristor crossbars. A predictive closed‐form distributed line‐resistance model, combined with a fourmatrix geometry‐averaging algorithm, eliminates these parasitic limitations.
Davide Rossetti   +5 more
wiley   +1 more source

Learning with Box Kernels

IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011
Supervised examples and prior knowledge on regions of the input space have been profitably integrated in kernel machines to improve the performance of classifiers in different real-world contexts. The proposed solutions, which rely on the unified supervision of points and sets, have been mostly based on specific optimization schemes in which, as usual,
Melacci, Stefano, Gori, Marco
openaire   +7 more sources

Hierarchical kernels in deep kernel learning

J. Mach. Learn. Res., 2023
Summary: Kernel methods are built upon the mathematical theory of reproducing kernels and reproducing kernel Hilbert spaces. They enjoy good interpretability thanks to the solid mathematical foundation. Recently, motivated by deep neural networks in deep learning, which construct learning functions by successive compositions of activation functions and
Wentao Huang, Houbao Lu, Haizhang Zhang
openaire   +2 more sources

Learning by Kernel Polarization

Neural Computation, 2005
Kernels are key components of pattern recognition mechanisms. We propose a universal kernel optimality criterion, which is independent of the classifier to be used. Defining data polarization as a process by which points of different classes are driven to geometrically opposite locations in a confined domain, we propose selecting the kernel parameter ...
openaire   +2 more sources

Home - About - Disclaimer - Privacy