@inproceedings{5f7fa435e425418c8ca0d171c5bbc49c,
title = "IoT-XPLAIN: Intelligent Fault Detection and Natural Language Explanation for IoT Systems Using ML and LLMs",
abstract = "Modern IoT deployments face challenges in maintaining fault resilience due to limited observability, dynamic conditions, and constrained edge resources. Traditional detection methods often lack flexibility and explainability. This paper introduces IoT-XPLAIN, a hybrid framework integrating Random Forest classification with large language models (LLMs) for automated anomaly detection and human-readable fault explanations. Trained on TON\_IoT telemetry with DNSlayer metadata, the model detects transport-and application-level anomalies, including synthetic DNS faults. Upon detection, GPT40 generates summaries of fault patterns, causes, and recovery steps. IoT-XPLAIN enables interpretable, efficient fault analysis for smart manufacturing, remote monitoring, and fault-aware edge infrastructures.",
keywords = "Anomaly Detection, DNS Fault Injection, Explainable AI, Fault Detection, GPT-4o, Internet of Things (IoT), Random Forest",
author = "Harshit Vashisht and Singh, \{Sunil K.\} and Sudhakar Kumar and Nipun Singh and Varsha Arya and Chui, \{Kwok Tai\} and Gupta, \{Brij B.\}",
note = "Publisher Copyright: {\textcopyright} 2026 Global IT Research Institute - GIRI.; 28th International Conference on Advanced Communications Technology, ICACT 2026 ; Conference date: 08-02-2026 Through 11-02-2026",
year = "2026",
doi = "10.23919/ICACT68090.2026.11431335",
language = "English",
series = "International Conference on Advanced Communication Technology, ICACT",
pages = "66--71",
booktitle = "28th International Conference on Advanced Communications Technology",
}