Artificial Intelligence Powered Memristive Average Value Calculator and its Real-time Applications

Abstract

Recently, disasters due to floods based on extreme rainfall have resulted in over 2 million deaths and US$ 3.64 trillion loss of property over the last five decades globally. On the other side, researchers have been increasing interest in employing large language models (LLMs) to generate solutions for numerous real-time applications. This work proposes memristor-based electronic circuits for flood alert indication developed through LLM after training through an example. Initially, an average value calculator was designed to record the rainfall associated with four consecutive days. Importantly, a memristor-based electronic circuit was developed to provide extreme rainfall alert indication during heavy rain. Most importantly, the employment of LLM was advantageous in terms of ease and reduction in code complexity by 98.94%. Furthermore, the power consumed by our proposed circuit resulted in ~ 99.9% improvement, whereas the circuit complexity was reduced by ~ 99% when compared to their existing transistor-based counterparts.