A Generic Layer Pruning Method for Signal Modulation Recognition Deep Learning Models
Zhejiang University of Technology（浙江工业大学）；Binjiang Institute of Artificial Intelligence, Zhejiang University of Technology（浙江工业大学滨江人工智能研究院）；Institute of Cyberspace Security, College of Information Engineering, Zhejiang University of Technology（浙江工业大学信息工程学院网络空间安全研究院）；Institute of Computing Technology, Chinese Academy of Sciences（中国科学院计算技术研究所）；Harbin Engineering University（哈尔滨工程大学）；PCL Research Center of Networks and Communications, Peng Cheng Laboratory（鹏城实验室PCL网络与通信研究中心）；Utron Technology Company Ltd.（优创科技有限公司）；Science and Technology on Communication Information Security Control Laboratory（通信信息安全控制重点实验室）
Yao Lu, Yutao Zhu, Yuqi Li, Dongwei Xu, Yun Lin, Qi Xuan, Xiaoniu Yang
With the successful application of deep learning in communications systems, deep neural networks are becoming the preferred method for signal classification. Although these models yield impressive results, they often come with high computational complexity and large model sizes, which hinders their practical deployment in communication systems. To address this challenge, we propose a novel layer pruning method. Specifically, we decompose the model into several consecutive blocks, each containing consecutive layers with similar semantics. Then, we identify layers that need to be preserved within each block based on their contribution. Finally, we reassemble the pruned blocks and fine-tune the compact model. Extensive experiments on five datasets demonstrate the efficiency and effectiveness of our method over a variety of state-of-the-art baselines, including layer pruning and channel pruning methods.
FCOS: A Two-Stage Recoverable Model Pruning Framework for Automatic Modulation Recognition
Institute of Cyberspace Security, College of Information Engineering, Zhejiang University of Technology（信息工程学院网络安全研究所，浙江工业大学）；Binjiang Institute of Artificial Intelligence, Zhejiang University of Technology（滨江人工智能研究院，浙江工业大学）；Research Center of Electromagnetic SpaceSecurity, Binjiang Institute of Artificial Intelligence, ZJUT（电磁空间安全研究室，滨江人工智能研究院，ZJUT）；College of Information and Communication Engineering, Harbin Engineering University（信息与通信工程学院，哈尔滨工程大学）
An Efficient Deep Learning Model for Automatic Modulation Recognition Based on Parameter Estimation and Transformation
School of Information and Communication Engineering, University of Electronic Science and Technology of China（电子科技大学信息与通信工程学院）；Yangtze Delta Region Institute (Huzhou), University of Electronic Science and Technology of China（电子科技大学长三角研究院（湖州））
DUSE: A Data Expansion Framework for Low-resource Automatic Modulation Recognition based on Active Learning
Institute of Cyberspace Security, College of Information Engineering, Zhejiang University of Technology（网络安全研究所，信息工程学院，浙江工业大学）；Binjiang Institute of Artificial Intelligence, Zhejiang University of Technology（滨江人工智能研究院，浙江工业大学）；Centre for Frontier AI Research, Agency for Science, Technology and Research（前沿人工智能研究中心，科技研究局）；College of Information and Communication Engineering, Harbin Engineering University（信息与通信工程学院，哈尔滨工程大学）；College of Telecommunications and Information Engineering, Nanjing University of Posts and Telecommunications（电信与信息工程学院，南京邮电大学）
Play and Prune: Adaptive Filter Pruning for Deep Model Compression
IIT Kanpur（印度理工学院坎普尔分校）
虽然卷积神经网络（CNN）在各种分类/识别任务上取得了令人印象深刻的性能，但它们通常由大量参数组成。这导致显著的存储器需求以及计算开销。因此，越来越需要用于压缩基于CNN的模型的滤波器级修剪方法，其不仅减少参数的总数，而且还减少总体计算。我们提出了一个新的最小-最大框架，用于CNN的过滤器级修剪。我们的框架称为Play and Prune（PP），它通过自适应修剪率联合修剪和微调CNN模型参数，同时保持模型的预测性能。我们的框架由两个模块组成：（1）自适应过滤器修剪（AFP）模块，它最大限度地减少模型中的过滤器数量;（2）修剪速率控制器（PRC）模块，它最大限度地提高修剪过程中的准确性。此外，与大多数以前的方法不同，我们的方法允许直接指定所需的容错性，而不是修剪水平。我们的压缩模型可以在运行时部署，而不需要任何特殊的库或硬件。我们的方法将VGG-16的参数数量减少了17.5倍，FLOPS数量减少了6.43倍，并且没有损失准确性，显著优于其他最先进的过滤器修剪方法。
Enhancing Automatic Modulation Recognition for IoT Applications Using Transformers
The University of North Carolina at Charlotte（北卡罗来纳大学夏洛特分校）；Kostas Research Institute at Northeastern University（东北大学科斯塔斯研究所）；Northeastern University（东北大学）
MDM: Advancing Multi-Domain Distribution Matching for Automatic Modulation Recognition Dataset Synthesis
College of Information Engineering, Zhejiang University of Technology（浙江工业大学信息工程学院）；Institute of Cyberspace Security, Zhejiang University of Technology（浙江工业大学网络空间安全研究院）；Institute of China Electronics Technology Group Corporation 36th Research Institute（中国电子科技集团公司第三十六研究所）；College of Information And Communication Engineering, Harbin Engineering University（哈尔滨工程大学信息与通信工程学院）；National Key Laboratory of Electromagnetic Space Security（电磁空间安全全国重点实验室）
Low-Interception Waveform: To Prevent the Recognition of Spectrum Waveform Modulation via Adversarial Examples
Qian Xuesen Laboratory of Space Technology, China Academy of Space Technology（中国空间技术研究院钱学森空间技术实验室）；School of Mathematics and Computational Science, Xiangtan University（湘潭大学数学与计算科学学院）
Rethinking Impersonation and Dodging Attacks on Face Recognition Systems
Huazhong University of Science and Technology（华中科技大学）；Shanghai Jiao Tong University（上海交通大学）；Shanghai Shizhuang Information Technology Co., Ltd（上海石庄信息技术有限公司）；La Trobe University（拉筹伯大学）
AI/ML-Based Automatic Modulation Recognition: Recent Trends and Future Possibilities
Data Science and Analytics Institute, University of Oklahoma（俄克拉荷马大学数据科学与分析研究所）；University of New Mexico（新墨西哥大学）；Labcorp Drug Development（Labcorp药物研发公司）；Samsung Research America（三星美国研究院）；Maxwave LLC（Maxwave有限责任公司）
该研究在统一数据集、硬件和准确率指标下复现并比较多种基于 AI/ML 的 AMR 模型，结合参数复杂度评估性能，并指出架构、数据处理与训练策略的改进方向。
Ultralight Signal Classification Model for Automatic Modulation Recognition
Multiverse Computing（多元宇宙计算公司）；DAS Photonics（DAS光子学公司）；DIPC (Donostia International Physics Center)（多诺斯蒂亚国际物理中心（DIPC））；Ikerbasque Foundation for Science（伊克尔巴斯克科学基金会）
Deep Learning for Multi-Antenna Modulation Recognition of Radio Signals
College of Electrical and Information Engineering, Quzhou University, Quzhou 324000, China（庆州大学电气与信息工程学院）；National Key Laboratory of Electromagnetic Space Security, Jiaxing 314033, China（电磁空间安全国家重点实验室）；Institute of Cyberspace Security（网络空间安全研究院）；College of Information Engineering, Zhejiang University of Technology, Hangzhou 310023, China（浙江工业大学信息工程学院）
MCLRL: A Multi-Domain Contrastive Learning with Reinforcement Learning Framework for Few-Shot Modulation Recognition
Institute of Cyberspace Security, College of Information Engineering, Zhejiang University of Technology（浙江工业大学信息工程学院网络空间安全研究所）；Binjiang Institute of Artificial Intelligence, Zhejiang University of Technology（浙江工业大学滨江人工智能研究院）；College of Information and Communication Engineering, Harbin Engineering University（哈尔滨工程大学信息与通信工程学院）
