结合二值化神经网络与知识蒸馏的轻量型水声目标识别算法
CSTR:
作者:
作者单位:

北京理工大学 信息与电子学院,北京 100081

作者简介:

通讯作者:

E-mail: hrzlpk2015@gmail.com.

中图分类号:

TP391

基金项目:

国家自然科学基金项目(62301041).


A lightweight underwater acoustic target recognition algorithm combined with binarized neural networks and knowledge distillation
Author:
Affiliation:

School of Information and Electronics,Beijing Institute of Technology,Beijing 100081,China

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    基于深度学习的水下声学目标识别算法在水下平台部署时,通常面临计算资源短缺和水下声学环境复杂多变的挑战,提出一种结合二值化神经网络与知识蒸馏的轻量型水声目标识别算法(DSBNN_KD),旨在通过深度可分离卷积和权重参数二值化等手段实现模型的压缩和优化加速.同时,利用知识蒸馏技术将高性能高复杂度的教师模型的知识转移到轻量级学生模型上,从而缓解极端量化带来的性能损失,并确保模型的泛化性能.对DSBNN_KD的表现在两个实测水声数据集上进行全面评估,实验结果表明,相比当前主流的全精度轻量化模型,所提出DSBNN_KD在模型参数量、模型部署尺寸和计算量等方面均展现出显著的优势,同时在知识蒸馏技术的辅助下,量化后的模型依然可以保持与全精度模型接近的性能.

    Abstract:

    Underwater acoustic target recognition (UATR) algorithms based on deep learning often face the challenges of scarce computing resources and the complex and variable underwater acoustic environment when deployed on underwater platforms. Therefore, this paper proposes a lightweight UATR algorithm depthwise separable binarized neural network with knowledge distillation(DSBNN_KD), so as to achieve model compression and optimized acceleration through means of depth-separable convolution and weight parameter binarization. Meanwhile, the KD is utilized to transfer knowledge from high-performance, high-complexity teacher models to lightweight student models, thereby mitigating the performance loss caused by extreme quantization and ensuring the model's generalization performance. The performance of the DSBNN_KD is comprehensively evaluated on two observed underwater acoustic datasets. The experimental results indicate that, compared to current mainstream full-precision lightweight models, the proposed DSBNN_KD shows significant advantages in terms of model parameter volume, model deployment size, and computational load. With the assistance of the KD, the quantized model can still maintain performance close to that of full-precision models.

    参考文献
    相似文献
    引证文献
引用本文

崔翰林,褚晓晖,徐立军,等.结合二值化神经网络与知识蒸馏的轻量型水声目标识别算法[J].控制与决策,2025,40(1):128-136

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:
  • 最后修改日期:
  • 录用日期:
  • 在线发布日期: 2024-12-12
  • 出版日期: 2025-01-20
文章二维码