SurveyX: Academic Survey Automation via Large Language Models
Xun Liang, Jiawei Yang, Yezhaohui Wang, Chen Tang, Zifan Zheng, Simin Niu, Shichao Song, Hanyu Wang, Bo Tang, Feiyu Xiong, Keming Mao, Zhiyu li
2025-02-24
Summary
This paper talks about SurveyX, a new system that uses artificial intelligence to automatically create academic surveys, which are like detailed reports summarizing research in a specific field.
What's the problem?
Researchers are struggling to keep up with the huge amount of new academic papers being published every day. Writing surveys manually takes a lot of time and effort, and current AI systems for creating surveys have limitations like not being able to handle long texts or provide in-depth discussions.
What's the solution?
The researchers created SurveyX, which works in two main steps: preparation and generation. It uses online searches to find relevant information, organizes it using a method called AttributeTree, and then uses AI to write and polish the survey. SurveyX can understand and summarize long texts better than previous systems, and it can create more detailed and accurate surveys.
Why it matters?
This matters because it could save researchers a lot of time and help them stay up-to-date with new discoveries in their fields. SurveyX performs almost as well as human experts in creating these surveys, which means it could be a valuable tool for scientists and students who need to quickly understand the current state of research on a topic. This could speed up scientific progress and make it easier for people to learn about complex subjects.
Abstract
Large Language Models (LLMs) have demonstrated exceptional comprehension capabilities and a vast knowledge base, suggesting that LLMs can serve as efficient tools for automated survey generation. However, recent research related to automated survey generation remains constrained by some critical limitations like finite context window, lack of in-depth content discussion, and absence of systematic evaluation frameworks. Inspired by human writing processes, we propose SurveyX, an efficient and organized system for automated survey generation that decomposes the survey composing process into two phases: the Preparation and Generation phases. By innovatively introducing online reference retrieval, a pre-processing method called AttributeTree, and a re-polishing process, SurveyX significantly enhances the efficacy of survey composition. Experimental evaluation results show that SurveyX outperforms existing automated survey generation systems in content quality (0.259 improvement) and citation quality (1.76 enhancement), approaching human expert performance across multiple evaluation dimensions. Examples of surveys generated by SurveyX are available on www.surveyx.cn