Algorithm Skeleton Mining Based on Optimized Pre-training Mechanisms in Survey Paper

Analyzing the quality of an algorithm evolution process and constructing the knowledge body is a crucial research task with immense value. Thus, it becomes imperative to develop methods that can accurately predict the evolution of these algorithms. The paper creates a structured algorithm skeleton in survey papers to analyze lengthy texts effectively, focusing on key information and its context. It merges the reasoning abilities of extensive language models with optimized structured extraction techniques to enhance the understanding of essential information in lengthy texts. This study aims to facilitate the algorithm inventors to use the proper and high-quality way to evolve algorithms. Our research equips algorithm creators with tools and knowledge for the efficient development of high-quality evolution algorithms, contributing to the field’s advancement and real-world impact.