![[2602.22488] Explainability-Aware Evaluation of Transfer Learning Models for IoT DDoS Detection Under Resource Constraints](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.22488] Explainability-Aware Evaluation of Transfer Learning Models for IoT DDoS Detection Under Resource Constraints
Summary
This article evaluates transfer learning models for IoT DDoS detection, focusing on explainability and resource constraints. It analyzes various CNN architectures to determine their performance and interpretability in real-world applications.
Why It Matters
As IoT devices proliferate, they become increasingly vulnerable to DDoS attacks. Understanding the reliability and interpretability of detection models is crucial for deploying effective security measures in resource-constrained environments. This study provides insights into selecting suitable models that balance performance and explainability.
Key Takeaways
- DenseNet and MobileNet architectures excel in DDoS detection performance and reliability.
- The study emphasizes the need for explainability in AI models used for security applications.
- MobileNetV3 offers a favorable latency-accuracy trade-off for fog-level deployment.
- Integrating performance metrics with interpretability assessments enhances model selection.
- The findings guide practitioners in choosing deep learning models that meet operational constraints.
Computer Science > Cryptography and Security arXiv:2602.22488 (cs) [Submitted on 25 Feb 2026] Title:Explainability-Aware Evaluation of Transfer Learning Models for IoT DDoS Detection Under Resource Constraints Authors:Nelly Elsayed View a PDF of the paper titled Explainability-Aware Evaluation of Transfer Learning Models for IoT DDoS Detection Under Resource Constraints, by Nelly Elsayed View PDF HTML (experimental) Abstract:Distributed denial-of-service (DDoS) attacks threaten the availability of Internet of Things (IoT) infrastructures, particularly under resource-constrained deployment conditions. Although transfer learning models have shown promising detection accuracy, their reliability, computational feasibility, and interpretability in operational environments remain insufficiently explored. This study presents an explainability-aware empirical evaluation of seven pre-trained convolutional neural network architectures for multi-class IoT DDoS detection using the CICDDoS2019 dataset and an image-based traffic representation. The analysis integrates performance metrics, reliability-oriented statistics (MCC, Youden Index, confidence intervals), latency and training cost assessment, and interpretability evaluation using Grad-CAM and SHAP. Results indicate that DenseNet and MobileNet-based architectures achieve strong detection performance while demonstrating superior reliability and compact, class-consistent attribution patterns. DenseNet169 offers the strongest reliabi...
