Reproduce “SurveyAgent: A Conversational System for Personalized and Efficient Research Survey” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Xintao Wang, Jiangjie Chen, Nianqi Li, Lida Chen, Xinfeng Yuan, Wei Shi, Xuyang Ge, Rui Xu and 1 more
Fudan University researchers developed SURVEYAGENT, a conversational AI system that assists researchers with literature surveys by integrating knowledge management, personalized recommendations, and query answering. The system leverages LLM-based semantic filtering to enhance recommendation relevance and a ReAct-style agent framework for robust action planning, achieving an average F1-score of 93% for action selection and outperforming baselines in paper recommendation, with a notably higher Top-K Hit Ratio (e.g., 0.205 vs 0.194 for 9 seeds) compared to existing methods.
Key insights
- A unified conversational AI agent can effectively bridge the gaps between disparate research support tools, offering a holistic user experience.
- Combining traditional recommendation systems with LLM-based semantic filtering significantly improves the relevance and quality of paper suggestions.
- The ReAct framework enables robust multi-action planning for complex research queries, demonstrating flexibility and effectiveness beyond rigid predefined sequences.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.