What is SurveyGO
SurveyGO (Juanji) is an open-source tool by the Tsinghua and ModelBest team.AI paper writing tool. It supports quickly generating well-structured and content-rich survey articles based on the paper topic and description entered by the user, helping researchers and students efficiently complete the literature review section. SurveyGO's core technology, LLMxMapReduce-V2, draws on the principles of convolutional neural networks and aggregates multiple references based on a text convolution algorithm, avoiding the problem of information loss in traditional methods. The generated articles are logically rigorous, reliably cited, and rich in length and detail. Users only need to simply fill in information and submit a topic, and high-quality survey documents can be generated in the background, greatly improving the efficiency of research writing.
Main Features of SurveyGO
- Quickly generate survey articles: After the user enters the paper topic and description, it quickly generates a well-structured and content-rich survey, supporting the output of long texts of tens of thousands of words.
- User interaction: Based on like and comment functions, users provide feedback on writing needs and generated articles, and the system optimizes content generation according to user feedback.
- Multilingual support: Provides options for generating surveys in Chinese and English to meet the needs of different users.
- User-friendly interface: Simple to operate, supporting normal and professional modes, making it convenient for users to submit topics and view results.
Technical Principles of SurveyGO
- LLMxMapReduce-V2 technology: Uses a text convolution algorithm to aggregate multiple references. Similar to the application of convolutional neural networks (CNN) in image processing, it gradually abstracts local information into high-level global representations. The model first focuses on some cited articles (local information), and based on multi-layer convolution operations, integrates local information into more comprehensive structured information, such as the paragraph structure and themes of the article. An information entropy estimation module is introduced to guide the convolution process, ensuring that the scaling process during testing continuously improves the information content of the results.
- Addressing context length limitations: Traditional retrieval-augmented generation (RAG) methods generate content based on retrieving text fragments most relevant to the query, and easily ignore some important content that is related to the topic but cannot directly match semantic similarity. Based on aggregating multiple documents rather than simple retrieval, it makes full use of the information in all references and avoids information loss.
- High-quality content generation: Based on multi-layer convolution operations, it extracts key information from the literature and integrates it into structured output, ensuring the logic and coherence of the generated article. It automatically cites relevant literature and materials, ensuring the credibility and professionalism of the generated content.
- Evaluation benchmark: To scientifically evaluate the quality of generated articles, it creates a high-quality survey writing benchmark, SurveyEval, which is the first scalable evaluation benchmark in the field of computer science that combines surveys with complete references.
How to Use SurveyGO
- Visit the website: Visit SurveyGO'sofficial website. Complete registration and login according to the prompts.
- Choose a mode：
- Normal mode: Suitable for beginners or users with low requirements for generated content.
- Professional mode: Suitable for users with higher requirements for survey content.
- Fill in information：
- Paper title: Enter the topic for which you want to generate a survey.
- Paper Description: Briefly describe your research direction or the content you need to focus on.
- Language Selection: Select the language for generating the article (Chinese or English).
- Submit Topic: After filling in the information and submitting, the system automatically starts generating the survey article.
- View Generation Results: After generation is complete, view the generated article.
SurveyGO Project Address
- Project Official Website：https://surveygo.thunlp.org/
- GitHub Repository：https://github.com/thunlp/LLMxMapReduce
- HuggingFace Model Hub：https://huggingface.co/datasets/R0k1e/SurveyEval
- arXiv Technical Paper：https://arxiv.org/pdf/2504.05732
SurveyGO Application Scenarios
- Research Paper Writing: Helps researchers quickly generate the literature review section, saving time on searching for materials and organizing ideas.
- Academic Report Preparation: When preparing academic reports, quickly provides background information and research progress on related topics, helping users quickly build a report framework.
- Course Paper Writing: For students, it is a powerful tool for writing the literature review section of course papers, providing a clear logical structure and rich reference content.
- Industry Research and Analysis: Used to generate research reports for specific industries, analyze market trends, technological developments, etc., providing references for business decisions.
- Learning and Knowledge Expansion: Quickly understand the core viewpoints and research trends of new fields, and expand knowledge.