Home / Current Issue / Paper 1707330
Hybrid Machine Learning Frameworks: Bridging Quantum and Classical Computing for IoT Advancements
Subject area: Science,Engineering and Technology · Area of research: Internet of Things
Abstract
one of the maximum important use instances for deep?learning is image class. the appearance?of quantum technology has extended studies into quantum neural networks (QNNs). In conventional deep getting to know-based totally picture type, the capabilities of the photograph are extracted using a convolutional neural community (CNN)?and choice barriers are defined the usage of a multi-layer perceptron (MLP) network. Conversely, parameterized quantum?circuits can generate complex boundaries on selections and extract rich capabilities from images. This study proposed a hybrid QNN (H-QNN) model in binary picture class scenario to advantage from each QNN and?quantum computing. Our H-QNN model is distinctly efficient for computation on nosier intermediate-scale quantum (NISQ) devices, which are the?front-give up for quantum computing packages nowadays. this is accomplished by way of using a tensor product country of a small, -qubit quantum circuit to?be paired with a classical convolutional architecture. The?proposed H-QNN version can achieve 90.1% accuracy on binary image datasets, which substantially improves the classification accuracy. greater importantly, the proposed H-QNN and?the baseline CNN models are substantially evaluated at the image retrieval tasks as nicely. Quantitative consequences received show the generalisation of our H-QNN?for the downstream image retrieval tasks. by means of addressing the overfitting hassle for small datasets, our version is a valuable resource?for real-world applications.
Keywords
Quantum Convolutional Neural Networks, Hybrid Quantum?Classical Neural Networks, Image Retrieval, Classification, and Quantum Machine Learning.
How to cite this paper
@article{1707330,
author = {Amjad Khan, Ashish Kumar Pandey},
title = {Hybrid Machine Learning Frameworks: Bridging Quantum and Classical Computing for IoT Advancements},
journal = {Iconic Research And Engineering Journals},
year = {2025},
volume = {8},
number = {10},
pages = {425-436},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1707330.pdf},
abstract = {one of the maximum important use instances for deep?learning is image class. the appearance?of quantum technology has extended studies into quantum neural networks (QNNs). In conventional deep getting to know-based totally picture type, the capabilities of the photograph are extracted using a convolutional neural community (CNN)?and choice barriers are defined the usage of a multi-layer perceptron (MLP) network. Conversely, parameterized quantum?circuits can generate complex boundaries on selections and extract rich capabilities from images. This study proposed a hybrid QNN (H-QNN) model in binary picture class scenario to advantage from each QNN and?quantum computing. Our H-QNN model is distinctly efficient for computation on nosier intermediate-scale quantum (NISQ) devices, which are the?front-give up for quantum computing packages nowadays. this is accomplished by way of using a tensor product country of a small, -qubit quantum circuit to?be paired with a classical convolutional architecture. The?proposed H-QNN version can achieve 90.1% accuracy on binary image datasets, which substantially improves the classification accuracy. greater importantly, the proposed H-QNN and?the baseline CNN models are substantially evaluated at the image retrieval tasks as nicely. Quantitative consequences received show the generalisation of our H-QNN?for the downstream image retrieval tasks. by means of addressing the overfitting hassle for small datasets, our version is a valuable resource?for real-world applications.},
keywords = {Quantum Convolutional Neural Networks, Hybrid Quantum?Classical Neural Networks, Image Retrieval, Classification, and Quantum Machine Learning.},
month = {April},
}