Development and Implementation of a Corrosion Inhibitor Chatbot Using Bidirectional Long Short-Term Memory

Abstract

This research delves into the intricate phenomenon of corrosion, a process entailing material degradation through chemical reactions with the environment, causing consequential losses across diverse sectors. In response, corrosion inhibitors are a proactive measure to counteract this deleterious impact. Despite their paramount significance, public awareness regarding corrosion and inhibitors remains limited, necessitating intensified educational efforts. The primary focus of this study is developing a Chatbot system designed to disseminate information on corrosion, inhibitors, and related topics. Employing the Machine Learning Life Cycle model, a deep learning approach, specifically the Bidirectional Long Short-Term Memory (BLSTM) architecture, is utilized to construct an optimized Chatbot model. Post-training evaluation of the BLSTM model reveals noteworthy performance metrics, including a remarkable 100% accuracy rate and a substantial 92% validation accuracy over 100 epochs. Training and validation losses are reported as 0.2292 and 0.9342, respectively. In conclusion, the BLSTM algorithm is an effective tool for training and enhancing Chatbot models, ensuring commendable corrosion awareness and inhibition performance.