New framework automates academic survey generation

Researchers have introduced DAS, a stateful agentic framework designed to automate the creation of academic surveys. This system separates paper analysis from manuscript construction, utilizing a dynamically updated metadata lake of approximately two million papers. DAS employs agents to manage literature, organization, writing, and finalization states, incorporating a semantic review process for efficient updates. A new benchmark, DAS-Bench, and evaluation metric, DAS-Eval, were also developed to assess the quality of generated surveys across various criteria. AI

IMPACT This framework could significantly streamline academic research by automating literature review and survey generation.

RANK_REASON Academic paper introducing a new framework and benchmark. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• alphaXiv
• AutoSurveyGPT: GPT-Enhanced Automated Literature Discovery
• CatalyzeX Code Finder for Papers
• Connected Papers
• DagsHub
• DAS
• DAS-2M
• DAS-Bench
• DAS-Eval
• Gotit.pub
• Hugging Face
• Litmaps
• Naive RAG
• ScienceCast
• scite Smart Citations

• paper
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →