{"title":"多访问边缘计算中CNN推理的动态语义压缩:基于图强化学习的自编码器","authors":"Nan Li;Alexandros Iosifidis;Qi Zhang","doi":"10.1109/TWC.2024.3518399","DOIUrl":null,"url":null,"abstract":"This paper studies the computational offloading of CNN inference in dynamic multi-access edge computing (MEC) networks. To address the uncertainties in communication time and edge servers’ available capacity, we propose a novel semantic compression method, autoencoder-based CNN architecture (AECNN), for effective semantic extraction and compression in partial offloading. In the semantic encoder, we introduce a feature compression module based on the channel attention mechanism in CNNs, to compress intermediate data by selecting the most informative features. Additionally, to further reduce communication overhead, we leverage entropy encoding to remove the statistical redundancy in the compressed data. In the semantic decoder, we design a lightweight decoder to reconstruct the intermediate data through learning from the received compressed data to improve accuracy. To effectively trade-off communication, computation, and inference accuracy, we design a reward function and formulate the offloading problem of CNN inference as a maximization problem with the goal of maximizing the average inference accuracy and throughput over the long term. To address this maximization problem, we propose a graph reinforcement learning-based AECNN (GRL-AECNN) method, which outperforms existing works DROO-AECNN, GRL-BottleNet++ and GRL-DeepJSCC under different dynamic scenarios. This highlights the advantages of GRL-AECNN in offloading decision-making for CNN inference tasks in dynamic MEC.","PeriodicalId":13431,"journal":{"name":"IEEE Transactions on Wireless Communications","volume":"24 3","pages":"2157-2172"},"PeriodicalIF":8.9000,"publicationDate":"2024-12-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Dynamic Semantic Compression for CNN Inference in Multi-Access Edge Computing: A Graph Reinforcement Learning-Based Autoencoder\",\"authors\":\"Nan Li;Alexandros Iosifidis;Qi Zhang\",\"doi\":\"10.1109/TWC.2024.3518399\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"This paper studies the computational offloading of CNN inference in dynamic multi-access edge computing (MEC) networks. To address the uncertainties in communication time and edge servers’ available capacity, we propose a novel semantic compression method, autoencoder-based CNN architecture (AECNN), for effective semantic extraction and compression in partial offloading. In the semantic encoder, we introduce a feature compression module based on the channel attention mechanism in CNNs, to compress intermediate data by selecting the most informative features. Additionally, to further reduce communication overhead, we leverage entropy encoding to remove the statistical redundancy in the compressed data. In the semantic decoder, we design a lightweight decoder to reconstruct the intermediate data through learning from the received compressed data to improve accuracy. To effectively trade-off communication, computation, and inference accuracy, we design a reward function and formulate the offloading problem of CNN inference as a maximization problem with the goal of maximizing the average inference accuracy and throughput over the long term. To address this maximization problem, we propose a graph reinforcement learning-based AECNN (GRL-AECNN) method, which outperforms existing works DROO-AECNN, GRL-BottleNet++ and GRL-DeepJSCC under different dynamic scenarios. This highlights the advantages of GRL-AECNN in offloading decision-making for CNN inference tasks in dynamic MEC.\",\"PeriodicalId\":13431,\"journal\":{\"name\":\"IEEE Transactions on Wireless Communications\",\"volume\":\"24 3\",\"pages\":\"2157-2172\"},\"PeriodicalIF\":8.9000,\"publicationDate\":\"2024-12-23\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"IEEE Transactions on Wireless Communications\",\"FirstCategoryId\":\"94\",\"ListUrlMain\":\"https://ieeexplore.ieee.org/document/10812766/\",\"RegionNum\":1,\"RegionCategory\":\"计算机科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"ENGINEERING, ELECTRICAL & ELECTRONIC\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"IEEE Transactions on Wireless Communications","FirstCategoryId":"94","ListUrlMain":"https://ieeexplore.ieee.org/document/10812766/","RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENGINEERING, ELECTRICAL & ELECTRONIC","Score":null,"Total":0}
Dynamic Semantic Compression for CNN Inference in Multi-Access Edge Computing: A Graph Reinforcement Learning-Based Autoencoder
This paper studies the computational offloading of CNN inference in dynamic multi-access edge computing (MEC) networks. To address the uncertainties in communication time and edge servers’ available capacity, we propose a novel semantic compression method, autoencoder-based CNN architecture (AECNN), for effective semantic extraction and compression in partial offloading. In the semantic encoder, we introduce a feature compression module based on the channel attention mechanism in CNNs, to compress intermediate data by selecting the most informative features. Additionally, to further reduce communication overhead, we leverage entropy encoding to remove the statistical redundancy in the compressed data. In the semantic decoder, we design a lightweight decoder to reconstruct the intermediate data through learning from the received compressed data to improve accuracy. To effectively trade-off communication, computation, and inference accuracy, we design a reward function and formulate the offloading problem of CNN inference as a maximization problem with the goal of maximizing the average inference accuracy and throughput over the long term. To address this maximization problem, we propose a graph reinforcement learning-based AECNN (GRL-AECNN) method, which outperforms existing works DROO-AECNN, GRL-BottleNet++ and GRL-DeepJSCC under different dynamic scenarios. This highlights the advantages of GRL-AECNN in offloading decision-making for CNN inference tasks in dynamic MEC.
期刊介绍:
The IEEE Transactions on Wireless Communications is a prestigious publication that showcases cutting-edge advancements in wireless communications. It welcomes both theoretical and practical contributions in various areas. The scope of the Transactions encompasses a wide range of topics, including modulation and coding, detection and estimation, propagation and channel characterization, and diversity techniques. The journal also emphasizes the physical and link layer communication aspects of network architectures and protocols.
The journal is open to papers on specific topics or non-traditional topics related to specific application areas. This includes simulation tools and methodologies, orthogonal frequency division multiplexing, MIMO systems, and wireless over optical technologies.
Overall, the IEEE Transactions on Wireless Communications serves as a platform for high-quality manuscripts that push the boundaries of wireless communications and contribute to advancements in the field.