Search

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. [ABSTRACT FROM AUTHOR]

As fire safety issues receive increasing attention, the development of efficient flame-retardant materials has become particularly important. This study explores the preparation methods of nano-violet phosphorus (NVP) and nanocellulose (NCC) composite films and their superior performance in flame retardancy applications. NVP was successfully prepared using a liquid-phase exfoliation method and was then non-covalently self-assembled with NCC to form a novel composite film with good uniformity. The microstructure, mechanical properties, and flame-retardant performance of the composite film were comprehensively characterized. The results indicated that the NVP/NCC composite film with only 1 wt% NVP exhibited significantly enhanced self-extinguishing capability (<2 s) and a peak heat release rate (23.31 kW m−2) and total heat release (1.78 MJ m−2), indicating its effectiveness in delaying the combustion process under flame exposure. The introduction of NVP made the decomposition rate of the composite film slower and the decomposition behavior more stable in an aerobic environment. In addition, the presence of NVP delayed the dehydration process of NCC, and promoted the carbonization reaction, thereby reducing the release of H2O, thus forming a protective carbon layer and removing active free radicals, significantly improving the thermal stability and safety of the composite. This work provides important theoretical foundations and practical guidance for the development of new flame-retardant materials. [ABSTRACT FROM AUTHOR]