arXiv:2605.26166v1 Announce Type: cross 
Abstract: The rapid proliferation of Internet of Things (IoT) devices has created an urgent demand for adaptive, resource-efficient Intrusion Detection Systems (IDS) capable of handling dynamic and evolving cyber threats. This paper investigates AOC-IDS, a state-of-the-art autonomous online IDS published at IEEE INFOCOM 2024, which employs an Autoencoder (AE) with Cluster Repelling Contrastive (CRC) loss and an autonomous Gaussian-based decision module. We first successfully replicate AOC-IDS on the UNSW-NB15 benchmark, achieving 89.39% accuracy in close agreement with the published 89.19%. We then identify four key limitations: class imbalance, unreliable pseudo-label generation, limited generalization, and computational overhead for IoT deployment, and propose targeted improvements for each. Our XGBoost-BalSamp method achieves 95.45% accuracy on UNSW-NB15, a gain of 6.26% over the baseline. Our combined deep learning approach (PseudoFilter, MixupAug, and LiteAE) achieves a best-run accuracy of 90.88% (F1: 91.45%), surpassing the base paper while reducing model parameters by 55%.These results demonstrate that targeted improvements to AOC-IDS yield consistent accuracy gains while improving practical deployability on IoT edge devices.

يقدم المقال AOC-IDS، وهو نظام متقدم للكشف عن التسلل عبر الإنترنت مصمم لإنترنت الأشياء (IoT)، حيث حقق دقة بنسبة 89.39% على معيار UNSW-NB15. يتناول مشكلات حرجة مثل عدم توازن الفئات وتوليد تسميات زائفة غير موثوقة، مقترحًا تحسينات تزيد من الأداء إلى 95.45% من الدقة.

El artículo presenta AOC-IDS, un sistema de detección de intrusiones en línea autónomo avanzado diseñado para el Internet de las Cosas (IoT), logrando un 89.39% de precisión en el benchmark UNSW-NB15. Aborda problemas críticos como el desequilibrio de clases y la generación de pseudoetiquetas poco confiables, proponiendo mejoras que aumentan el rendimiento al 95.45% de precisión.

L'article présente AOC-IDS, un système de détection d'intrusions en ligne autonome avancé conçu pour l'Internet des objets (IoT), atteignant 89,39 % de précision sur la référence UNSW-NB15. Il aborde des problèmes critiques tels que le déséquilibre des classes et la génération de pseudo-étiquettes peu fiables, proposant des améliorations qui augmentent la performance à 95,45 % de précision.

The paper presents AOC-IDS, an advanced autonomous online Intrusion Detection System (IDS) designed for the Internet of Things (IoT), achieving 89.39% accuracy on the UNSW-NB15 benchmark. It addresses critical issues such as class imbalance and unreliable pseudo-label generation, proposing improvements that enhance performance to 95.45% accuracy.