Zero-Shot Traffic Identification with Attribute and Graph-Based Representations for Edge Computing
With the proliferation of mobile terminals and the rapid growth of network applications, fine-grained traffic identification has become increasingly challenging. Methods based on machine learning and deep learning have achieved remarkable results, but they heavily rely on the distribution of trainin...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-01-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/25/2/545 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832587479981490176 |
|---|---|
| author | Zikui Lu Zixi Chang Mingshu He Luona Song |
| author_facet | Zikui Lu Zixi Chang Mingshu He Luona Song |
| author_sort | Zikui Lu |
| collection | DOAJ |
| description | With the proliferation of mobile terminals and the rapid growth of network applications, fine-grained traffic identification has become increasingly challenging. Methods based on machine learning and deep learning have achieved remarkable results, but they heavily rely on the distribution of training data, which makes them ineffective in handling unseen samples. In this paper, we propose AG-ZSL, a zero-shot learning framework based on traffic behavior and attribute representations for general encrypted traffic classification. AG-ZSL primarily learns two mapping functions: one that captures traffic behavior embeddings from burst-based traffic interaction graphs, and the other that learns attribute embeddings from traffic attribute descriptions. Then, the framework minimizes the distance between these embeddings within the shared feature space. The gradient rejection algorithm and K-Nearest Neighbors are introduced to implement a two-stage method for general traffic classification. Experimental results on IoT datasets demonstrate that AG-ZSL achieves exceptional performance in classifying both known and unknown traffic, highlighting its potential for enhancing secure and efficient traffic management at the network edge. |
| format | Article |
| id | doaj-art-1d8d714d179d477398c4275154d48598 |
| institution | Kabale University |
| issn | 1424-8220 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj-art-1d8d714d179d477398c4275154d485982025-01-24T13:49:17ZengMDPI AGSensors1424-82202025-01-0125254510.3390/s25020545Zero-Shot Traffic Identification with Attribute and Graph-Based Representations for Edge ComputingZikui Lu0Zixi Chang1Mingshu He2Luona Song3School of Electronic Engineering, Beijing University of Posts and Telecommunications, Beijing 100876, ChinaFedUni Information Engineering Institute, Hebei University of Science and Technology, Shijiazhuang 050018, ChinaSchool of Cyberspace Security, Beijing University of Posts and Telecommunications, Beijing 100876, ChinaSchool of Economics and Management, Beijing Information Science & Technology University, Beijing 100192, ChinaWith the proliferation of mobile terminals and the rapid growth of network applications, fine-grained traffic identification has become increasingly challenging. Methods based on machine learning and deep learning have achieved remarkable results, but they heavily rely on the distribution of training data, which makes them ineffective in handling unseen samples. In this paper, we propose AG-ZSL, a zero-shot learning framework based on traffic behavior and attribute representations for general encrypted traffic classification. AG-ZSL primarily learns two mapping functions: one that captures traffic behavior embeddings from burst-based traffic interaction graphs, and the other that learns attribute embeddings from traffic attribute descriptions. Then, the framework minimizes the distance between these embeddings within the shared feature space. The gradient rejection algorithm and K-Nearest Neighbors are introduced to implement a two-stage method for general traffic classification. Experimental results on IoT datasets demonstrate that AG-ZSL achieves exceptional performance in classifying both known and unknown traffic, highlighting its potential for enhancing secure and efficient traffic management at the network edge.https://www.mdpi.com/1424-8220/25/2/545edge computingtraffic classificationzero-shot learningtraffic representationdeep learninggraph neural networks |
| spellingShingle | Zikui Lu Zixi Chang Mingshu He Luona Song Zero-Shot Traffic Identification with Attribute and Graph-Based Representations for Edge Computing Sensors edge computing traffic classification zero-shot learning traffic representation deep learning graph neural networks |
| title | Zero-Shot Traffic Identification with Attribute and Graph-Based Representations for Edge Computing |
| title_full | Zero-Shot Traffic Identification with Attribute and Graph-Based Representations for Edge Computing |
| title_fullStr | Zero-Shot Traffic Identification with Attribute and Graph-Based Representations for Edge Computing |
| title_full_unstemmed | Zero-Shot Traffic Identification with Attribute and Graph-Based Representations for Edge Computing |
| title_short | Zero-Shot Traffic Identification with Attribute and Graph-Based Representations for Edge Computing |
| title_sort | zero shot traffic identification with attribute and graph based representations for edge computing |
| topic | edge computing traffic classification zero-shot learning traffic representation deep learning graph neural networks |
| url | https://www.mdpi.com/1424-8220/25/2/545 |
| work_keys_str_mv | AT zikuilu zeroshottrafficidentificationwithattributeandgraphbasedrepresentationsforedgecomputing AT zixichang zeroshottrafficidentificationwithattributeandgraphbasedrepresentationsforedgecomputing AT mingshuhe zeroshottrafficidentificationwithattributeandgraphbasedrepresentationsforedgecomputing AT luonasong zeroshottrafficidentificationwithattributeandgraphbasedrepresentationsforedgecomputing |