IoT-AMLHP: Aligned Multimodal Learning of Header-Payload Representations for Resource-Efficient Malicious IoT Traffic Classification

2025-04-21

AI Summary

To address coarse-grained flow-level features, high noise in raw-byte representations, and modality misalignment in malicious traffic classification for resource-constrained IoT devices, this paper proposes a header-payload-aligned multimodal lightweight representation and fusion framework. We innovatively introduce header/payload decoupled parsing and semantic alignment to construct structured packet-level multimodal representations. Furthermore, we design a gated-attention-based adaptive fusion mechanism integrated with depthwise separable convolutions, enabling robust cross-modal complementary modeling under ultra-low parameter budgets. Evaluated on multiple IoT traffic datasets, our model achieves >98.5% detection accuracy while reducing parameter count by 62% and inference latency by 57% compared to existing lightweight state-of-the-art methods—demonstrating significant improvements in both efficiency and effectiveness.

Technology Category

• Machine Learning: Multimodal Learning
• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Security and Privacy: Large-scale security measurements

Abstract

Traffic classification is crucial for securing Internet of Things (IoT) networks. Deep learning-based methods can autonomously extract latent patterns from massive network traffic, demonstrating significant potential for IoT traffic classification tasks. However, the limited computational and spatial resources of IoT devices pose challenges for deploying more complex deep learning models. Existing methods rely heavily on either flow-level features or raw packet byte features. Flow-level features often require inspecting entire or most of the traffic flow, leading to excessive resource consumption, while raw packet byte features fail to distinguish between headers and payloads, overlooking semantic differences and introducing noise from feature misalignment. Therefore, this paper proposes IoT-AMLHP, an aligned multimodal learning framework for resource-efficient malicious IoT traffic classification. Firstly, the framework constructs a packet-wise header-payload representation by parsing packet headers and payload bytes, resulting in an aligned and standardized multimodal traffic representation that enhances the characterization of heterogeneous IoT traffic. Subsequently, the traffic representation is fed into a resource-efficient neural network comprising a multimodal feature extraction module and a multimodal fusion module. The extraction module employs efficient depthwise separable convolutions to capture multi-scale features from different modalities while maintaining a lightweight architecture. The fusion module adaptively captures complementary features from different modalities and effectively fuses multimodal features.

Problem

• Limited IoT device resources hinder complex deep learning deployment
• Existing methods misalign headers and payloads causing noise
• Need efficient multimodal learning for accurate traffic classification

Innovation

• Aligned multimodal header-payload representation construction
• Lightweight depthwise separable convolutions feature extraction
• Adaptive multimodal fusion for complementary features