Automated Literature Review Using Large Language Models

The exponential growth of scholarly publications presents a challenge for researchers seeking to comprehensively review and synthesize existing knowledge. Traditional manual literature review methods often prove time-consuming and limited in scope. Therefore, there is a need for a more efficient and accurate approach to conducting a literature review that can streamline the process and reduce the workload for researchers. This research work introduces an innovative approach to a literature review of the scientific research papers from the Arxiv Repository on a particular topic of interest by leveraging the capabilities of large language models (LLMs) and pre-trained transformers. The proposed methodology involves utilizing advanced LLMs, such as Generative Pretrained Transformer 3 (GPT-3), to automate various stages of the literature review process. These models are trained on massive volumes of text data, allowing them to extract and evaluate information from a wide variety of scholarly sources. This approach enables researchers to quickly identify relevant articles and extract key insights while reducing the time and effort required for manual literature review. Experimental findings suggest that employing LLMs to automate literature reviews yields promising outcomes.