Low-Power AI Models for Personalized Healthcare and Bioinformatics Applications

Abstract

The increasing demand for personalized healthcare and bioinformatics applications necessitates efficient AI-driven solutions capable of operating on resource-constrained edge devices. Traditional deep learning models are often computationally intensive, making them unsuitable for real-time analysis in IoT-based healthcare systems. This research proposes the development of ultra-lightweight, energy-efficient AI models optimized for low-power wearable devices, biosensors, and mobile health (mHealth) applications. By leveraging model compression techniques such as quantization, pruning, and knowledge distillation, the study aims to reduce computational complexity while maintaining high accuracy in disease prediction, genomic analysis, and real-time patient monitoring. Additionally, a cross-layer optimization strategy will be explored to enhance the energy efficiency of AI-driven wireless transmission in body area networks (BANs). The proposed framework will be validated using real-world biomedical datasets, ensuring robust performance across varied physiological conditions. This research contributes to the advancement of low-power AI for next-generation bioinformatics, enabling scalable, real-time, and energy-efficient personalized healthcare solutions.