Transforming Mobile Sensor Data Security Using Finance AI and Cutting-Edge Deep Learning

EasyChair Preprint 14873

Abstract

In times when even mobile devices have become a common commodity, the integrity and authenticity of the data generated from those mobile sensors have now become cardinal concerns. Indeed, the much-increased utilisation of mobile sensors in many different applications, from health care monitoring to autonomous vehicles, requires the application of reliable methods of data authentication. This work discusses the improvement of authenticity with deep machine learning models for data obtained from mobile sensors. In this paper, we describe the limitations imposed by traditional authentication techniques and show how deep learning models overcome such limitations. Further, we demonstrate the superiority in performance of deep learning models by performing an in-depth anomaly detection-based analysis on a broad dataset of mobile sensor outputs to ensure integrity in the data. The results show that the models based on deep learning perform with a much higher degree of accuracy and reliability compared to traditional approaches. This will provide better security to the data transmitted by mobile sensors and open ways for further research in improving the mobile security protocols through advanced techniques of machine learning.

Keyphrases: Data Integrity, Deep Learning Models, Explainable AI (XAI), Hybrid Authentication Systems, IoT Security, LSTM networks, Mobile Sensor Data Authentication, Transfer Learning, adversarial attacks, real-time processing

@booklet{EasyChair:14873,
  author    = {Adeoye Qudus},
  title     = {Transforming Mobile Sensor Data Security Using Finance AI and Cutting-Edge Deep Learning},
  howpublished = {EasyChair Preprint 14873},
  year      = {EasyChair, 2024}}