Large Language Model and Artificial Intelligence Based Human Conversation Agent

Our research aims to investigate large language models for the development of an autonomous software system that can engage in conversations with humans. To achieve our objective, we leverage a machine learning transformer with large language model for the development of an autonomous conversation system. Our proposed methodology includes the usage of a quantized large language model and applies parameter efficient fine tuning using low rank adaptation. We train and fine-tune our model for the specific area of electrical solar energy systems. For this purpose, we create our own dataset for this particular domain. The dataset creation methodology involves selection of a set of books related to electrical solar energy systems and extraction of text. The text is then organized in questions and answers format. The model exhibits a low perplexity score suggesting good performance. Other metrics also indicate a fairly good performance of our model. To the best of our knowledge, our work is the first to develop and train a machine learning language model for an electrical solar energy system.