Results 61 to 70 of about 119 (119)

Disturbance‐Aware On‐Chip Training with Mitigation Schemes for Massively Parallel Computing in Analog Deep Learning Accelerator

Advanced Science, EarlyView.
This study proposes novel operational schemes to solve the write disturbance issues in oxide‐semiconductor and capacitor‐based synaptic devices (6T1C devices). These schemes effectively neutralize disturbances, enabling high‐performance on‐chip training of convolutional neural networks and reducing capacitor size over 100 times.
Jaehyeon Kang, Jongun Won, Narae Han, Sangjun Hong, Jee‐Eun Yang, Sangwook Kim, Sangbum Kim   +6 more
wiley   +1 more source

A Novel Dual‐Network Approach for Real‐Time Liveweight Estimation in Precision Livestock Management

Advanced Science, EarlyView.
A novel dual‐network framework is proposed for real‐time, non‐contact liveweight estimation of pigs. By extracting contour information instead of segmented images, the method achieves high accuracy (R2 = 0.993) and an exceptional speed of 1131.6 FPS. This approach enhances automation in livestock farming, providing a scalable and efficient solution for
Ximing Dong, Caiming Zhang, Peiyuan Wang, Dexuan Chen, Gang Jun Tu, Shuhong Zhao, Tao Xiang   +6 more
wiley   +1 more source

All‐Electrical Control of Spin Synapses for Neuromorphic Computing: Bridging Multi‐State Memory with Quantization for Efficient Neural Networks

Advanced Science, EarlyView.
This study develops three all‐electrically controlled, field‐free spintronic synapse devices for neuromorphic computing. The tilted anisotropy device achieves an 11‐state memory with minimal cycle‐to‐cycle variation (2%), enabling high‐accuracy neural network quantization (81.51% in ResNet‐18). These findings position spintronic synapses as a promising
Tzu‐Chuan Hsin, Chun‐Yi Lin, Po‐Chuan Wang, Chun Yang, Chi‐Feng Pai   +4 more
wiley   +1 more source

Hybrid Series of Carbon‐Vacancy Electrodes for Multi Chemical Vapors Diagnosis Using a Residual Multi‐Task Model

Advanced Science, EarlyView.
This study presents a graphene oxide‐based gas sensor array integrated with a deep learning framework for precise detection and classification of multiple gases. By combining advanced sensing materials and a 1D‐ResNet architecture, the system achieves high sensitivity and accuracy, demonstrating its potential for real‐time environmental monitoring ...
Tianci Liu, Yun Ji Hwang, Lu Zhang, Jongwoo Hong, Teajong Hwang, Seong Chan Jun   +5 more
wiley   +1 more source

Photonic Bayesian Neural Networks: Leveraging Programmable Noise for Robust and Uncertainty‐Aware Computing

Advanced Science, EarlyView.
This work introduces a novel photonic Bayesian neural network architecture that utilizes tunable photonic random number generators to independently control weight distribution parameters, significantly improving robustness, scalability, and accuracy while demonstrating practical applications in multimodal data processing and uncertainty‐aware computing.
Yangyang Zhuge, Zhihao Ren, Zian Xiao, Zixuan Zhang, Xinmiao Liu, Weixin Liu, Siyu Xu, Chong Pei Ho, Nanxi Li, Chengkuo Lee   +9 more
wiley   +1 more source

Explainable Deep Multilevel Attention Learning for Predicting Protein Carbonylation Sites

Advanced Science, EarlyView.
Selective carbonylation sites (SCANS) are conceptualized, designed, evaluated, and released. SCANS captures segment‐level, protein‐level, and residue embeddings features. It utilizes elaborate loss function to penalize cross‐predictions at the residue level.
Jian Zhang, Jingjing Qian, Pei Wang, Xuan Liu, Fuhao Zhang, Haiting Chai, Quan Zou   +6 more
wiley   +1 more source

Pushing the Frontiers: Artificial Intelligence (AI)‐Guided Programmable Concepts in Binary Self‐Assembly of Colloidal Nanoparticles

Advanced Science, EarlyView.
Colloidal nanoparticles self‐assembly advances towards intelligent, customized assembly through precise control of binary co‐assemblies. This review explores the evolution from monolithic to binary assemblies, highlighting how the AI‐guided programmable assembly approach has the potential to shift from passive assembly to active intelligent design.
Cancan Li, Lindong Ma, Zhenjie Xue, Xiao Li, Shan Zhu, Tie Wang   +5 more
wiley   +1 more source

Piezoresistive Effect: A New Concept for Hearing Aids

Advanced Science, EarlyView.
An MXene/PVA sound sensor utilizing piezoresistive effects has been proposed with a detection limit of 60 dB and a frequency response of 20–4000 Hz, which is comparable to the performance of commercial hearing aids. By combining the sensor with machine learning, the accuracy rate of voiceprint recognition of wildlife conservation can reach up to 99 ...
Mengyao Gao, Weijie Liu, Kun Chen, Huili Sun, Xiaoqing Liu, Haonan Xing, Huatang Wang, Benpeng Zhu, Haizhong Guo   +8 more
wiley   +1 more source

Artificial Olfactory System Enabled by Ultralow Chemical Sensing Variations of 1D SnO2 Nanoarchitectures

Advanced Science, EarlyView.
This study presents a highly reliable gas sensor platform featuring SnO2 nanonetworks functionalized with Au and Pd nanocatalysts. Enhanced stability and optimized performance enable over 99.5% classification accuracy in deep learning, even under extreme conditions.
Yun‐Haeng Cho, Dong‐Su Kim, Jung Hwan Seo, Jae Han Chung, Zion Park, Ki Chang Kwon, Jae‐Kwon Ko, Tae Won Ha, Jeong‐O Lee, Gyu‐Li Kim, Seong‐Jun Ro, Hyojung Kim, Chil‐Hyoung Lee, Kwangjae Lee, Young‐Seok Shim, Donghwi Cho   +15 more
wiley   +1 more source

Reversal of Stress‐Induced PIEZO1 Elevation with Mechanically Adapted Epicardial Patch for Myocardial Infarction Treatment

Advanced Science, EarlyView.
This work identifies fine‐tuning the expression of PIEZO1 as a critical molecular mechanism underlying the treatment of myocardial infarction by mechanically adapted cardiac patches, which can support the clinical translation of cardiac patch devices.
Yuwen Lu, Sibo Jiang, Ting Shen, Yuan Zhang, Chengbin He, Yun Gao, Liyin Shen, Qiao Jin, Yuting Zhao, Ping Liang, Chaochen Wang, Hongjie Hu, Jin He, Kaicheng Deng, Shuo Wang, Yunhe Chen, Jun Ling, Yang Zhu, Lenan Zhuang   +18 more
wiley   +1 more source