Scholarlink

<p>In recent years, machine learning (ML), supported by the surge in computational power, has been integrated into numerous research domains, providing vital solutions to a broad spectrum of challenges. However, the increasing volume and complexity of data utilized by ML systems have necessitated the exploration of new methods to quicken and refine the precision of existing ML algorithms. The utilization of satellite data in agriculture has grown extensively, providing numerous opportunities to improve different facets of agricultural management and productivity. Interestingly, the volume of data produced increases exponentially with time posing a challenge for classical ML computing methods. Quantum computers, capable of exploiting an exponentially vast quantum state space via quantum superposition and entanglement, are aptly suited to handle such data volumes. Therefore, the integration of classical ML algorithms with quantum computing, known as quantum machine learning (QML), holds significant promise in processing satellite data. In our research, we developed, trained, and evaluated novel hybrid classical-quantum multilayer neural networks, comparing their efficacy with conventional classical networks. We applied QML to two classification tasks, namely the classification of rice and cotton crops using satellite data obtained by Sentinel-1 and Sentinel-2 satellites. Our results show that integrating quantum neurons into classical multilayer neural networks substantially improves their performance.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Hybrid classical-quantum multilayer neural networks for monitoring agricultural activities using remote sensing data

  • 期刊论文

  • 发表日期：

    2025年1月4日

查看全文
  • Ioannis Liliopoulos,
  • Georgios D. Varsamis,
  • Kristin Milchanowski,
  • Rafael Martin‑Cuevas,
  • Konstantina Safouri,
  • Panagiotis Dimitrakis,
  • Ioannis G. Karafyllidis

摘要

In recent years, machine learning (ML), supported by the surge in computational power, has been integrated into numerous research domains, providing vital solutions to a broad spectrum of challenges. However, the increasing volume and complexity of data utilized by ML systems have necessitated the exploration of new methods to quicken and refine the precision of existing ML algorithms. The utilization of satellite data in agriculture has grown extensively, providing numerous opportunities to improve different facets of agricultural management and productivity. Interestingly, the volume of data produced increases exponentially with time posing a challenge for classical ML computing methods. Quantum computers, capable of exploiting an exponentially vast quantum state space via quantum superposition and entanglement, are aptly suited to handle such data volumes. Therefore, the integration of classical ML algorithms with quantum computing, known as quantum machine learning (QML), holds significant promise in processing satellite data. In our research, we developed, trained, and evaluated novel hybrid classical-quantum multilayer neural networks, comparing their efficacy with conventional classical networks. We applied QML to two classification tasks, namely the classification of rice and cotton crops using satellite data obtained by Sentinel-1 and Sentinel-2 satellites. Our results show that integrating quantum neurons into classical multilayer neural networks substantially improves their performance.