REVIEW | OPEN ACCESS Discovery 61, e7d1519 (2025) 1 of 9 Moderation of cyber attacks in IoTs using deep learning techniques Feanyi Stanly Nwokoro 1 , Shafi’I M Abdulhamid 2 , Kwaku Zacciah Adom-Oduro 3 , Anayo Chukwu Ikegwu 4 , Augustina Nebechi Nwatu 5 ABSTRACT The widespread use of IoT devices in the modern day has simplified our lives and elevated our everyday routines to a new level. IoT devices are connected to communicate and share data with gateways or access points (APs) for additional data processing. On the other hand, this makes cybersecurity and zero-day assaults in IoT networks more prevalent. Deep learning models and datasets used to identify fraudulent data in IoT environments have been reviewed in this research. In the context of the Internet of Things, we found that the Long Short-Term Memory (LSTM), Convolution Neural Network (CNN) and stacking auto-encoders improve the accuracy and precision of malicious packet detection. We undertook a thorough theoretical examination of deep learning datasets and models. Our research finding serves as paradigm to researchers as a technique to investigate IoT security and privacy challenges. Keywords: Internet of Things (IOTs); Access Points (APs); Long Short-Term Memory (LSTM); Convolution Neural Network (CNN); Deep Learning. 1. INTRODUCTION Over the past ten years, the Internet of Things (IoT) has advanced to transform the world using a new technology paradigm. It has the power to transform the lives of all people on the planet. IoT devices can communicate with one another and gather data from their surroundings to transmit to gateways, sensors, or any other device with internet access (Stoyanova et al., 2020). Smart homes, smart cars, smart grids, smart cities, smart agriculture, health care, surveillance, and supply chain management are just a few of the many uses for IoT today (Lee and Lee, 2015; Shin et al., 2019). Over 50 billion Internet of Things devices are expected to be online by the end of 2024 (Al- Garadi et al., 2020; Mekki et al., 2019). The proliferation of internet-connected devices has been supported by sensor downsizing, still, it has also raised privacy and security issues with the Internet of Things and left them open to numerous security breaches (Stoyanova et al., 2020; Tawalbeh et al., 2020). Discovery To Cite: Nwokoro FS, Abdulhamid SM, Adom-Oduro KZ, Ikegwu AC, Nwatu AN. Moderation of cyber attacks in IoTs using deep learning techniques. Discovery 2025; 61: e7d1519 doi: Author Affiliation: 1 Department of Computer Science, Rhema University, Nigeria 2 Department of Cyber Security, National Open University of Nigeria 3 Department of Computer Science, Technical University, Ghana 4 Department of Software Engineering, Veritas University, Nigeria 5 Department of Biotechnology, Alex Ekwueme Federal University (AE-FUNAI), Nigeria Peer-Review History Received: 04 October 2024 Reviewed & Revised: 08/October/2024 to 23/January/2025 Accepted: 27 January 2025 Published: 31 January 2025 Peer-Review Model External peer-review was done through double-blind method. Discovery pISSN 2278–5469; eISSN 2278–5450 © The Author(s) 2025. Open Access. This article is licensed under a Creative Commons Attribution License 4.0 (CC BY 4.0)., which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. DISCOVERY SCIENTIFIC SOCIETY