Results 231 to 240 of about 73,468 (309)

A Hybrid Transfer Learning Framework for Brain Tumor Diagnosis

Advanced Intelligent Systems, Volume 7, Issue 3, March 2025.
A novel hybrid transfer learning approach for brain tumor classification achieves 99.47% accuracy using magnetic resonance imaging (MRI) images. By combining image preprocessing, ensemble deep learning, and explainable artificial intelligence (XAI) techniques like gradient‐weighted class activation mapping and SHapley Additive exPlanations (SHAP), the ...
Sadia Islam Tonni, Md. Alif Sheakh, Mst. Sazia Tahosin, Md. Zahid Hasan, Taslima Ferdaus Shuva, Touhid Bhuiyan, Muhammad Ali Abdullah Almoyad, Nabil Anan Orka, Md. Tanvir Rahman, Risala Tasin Khan, M. Shamim Kaiser, Mohammad Ali Moni   +11 more
wiley   +1 more source

Convolutional neural networks: applications, challenges and future prospects in brain tumor research. [PDF]

Front Neurol
Zhang P, Yang Z.
europepmc   +1 more source

A Memristor‐Based In‐Memory Computing System‐on‐Chip with Efficient Depthwise Convolution

Advanced Intelligent Systems, EarlyView.
We present a memristor‐based in‐memory computing (IMC) architecture that enables efficient depthwise convolution (DWC) acceleration. Fabricated in a system‐on‐chip with crossbar arrays, the design improves memory utilization. Experimental validation demonstrates the first hardware acceleration of DWC in IMC, achieving a digital comparable inference ...
Wenhao Song, Fuxi Cai, Sangsu Park, Zhaoyan Li, Seonghun Lee, Mingyi Rao, Woohyung Choi, Seungin Shin, Yan Zhu, Sujin Chae, Myoungsub Kim, Mingche Wu, Hyejung Choi, Yuan Dao, Jaeyun Yi, Youngman Cho, Seon Yong Cha, Ning Ge, Qiangfei Xia, Soo Gil Kim, Miao Hu, J. Joshua Yang   +21 more
wiley   +1 more source

Federated training of spiking neural networks on edge hardware for audio processing. [PDF]

Front Neurosci
Kaimal SS, Jb A, Reka SS, Venugopal P.
europepmc   +1 more source

Explainable AI‐Driven Optimization of Electrode Activation Reduces Power Consumption While Preserving Object Recognition in Retinal Prostheses

Advanced Intelligent Systems, EarlyView.
Explainable artificial intelligence (XAI) guides selective electrode activation in retinal prostheses by emphasizing visually informative regions. XAI‐assisted phosphene generation maintains object recognition performance while significantly reducing stimulation power.
Sein Kim, Hamin Shim, Maesoon Im
wiley   +1 more source

Early retinal disease detection from fundus images using deep neural networks. [PDF]

Sci Rep
Hailu AA, Desta EA, Mulugeta A, Adugna FB, Salau AO, Almaw LA, Fentahun MB.   +6 more
europepmc   +1 more source

Resource‐Aware Contrastive Scattering Meta‐Learning for Efficient Few‐Shot Acoustic Anomaly Detection

Advanced Intelligent Systems, EarlyView.
This paper introduces a resource‐aware Contrastive Scattering Meta‐Learning (CSML) framework for acoustic anomaly detection. By leveraging training‐free wavelet scattering and metric‐based meta‐learning, the model achieves competitive performance with only 50 K learnable parameters—a 98% reduction compared to state‐of‐the‐art frameworks—enabling ...
Rami Zewail, Bassem Mokhtar
wiley   +1 more source

Application of deep learning in crop research: From genomics to phenomics. [PDF]

Plant Genome
Wu Z, Sun Y, Luo Q, Wei J, Cui J, Fang Y, Niu Y, Li Z, Wang X, Liu Z.   +9 more
europepmc   +1 more source

Identifying Dynamical Quantum Phase Transitions With a Migratable Quantum‐Classical Hybrid Neural Network

Advanced Intelligent Systems, EarlyView.
A hybrid quantum‐classical architecture is introduced to accurately identify dynamical quantum phase transitions from time‐evolved quantum states. The QCNN serves as a quantum dynamical feature extractor, while the classical network learns temporal correlations from a low‐dimensional readout sequence. The framework attains high accuracy, remains robust
Daili Li, Xin–Xin Yang, Xiaoting Wang, Wei Zhang   +3 more
wiley   +1 more source

Interpreting How Neural Networks Infer Scatterer Geometry from Echolocation Echoes

Advanced Intelligent Systems, EarlyView.
Neural networks enable echolocation‐based shape classification but remain difficult to interpret due to their black‐box nature. This work presents a feature‐importance metric to uncover the echo regions driving decisions in shape‐specialized networks.
Ganesh U. Patil, Hyung‐Suk Kwon, Bogdan‐Ioan Popa   +2 more
wiley   +1 more source