A new multi-agent system called ReviewPilot has been developed to automate the process of screening scholarly works and extracting data for systematic literature reviews. This system, which follows the PRISMA workflow, allows researchers to define criteria and schemas through dialogue and maintains a project memory for configurations. ReviewPilot has demonstrated superior performance compared to single-prompt LLMs and another system called AgentSLR in epidemiological reviews, achieving high F1 scores for both screening and extraction. AI
IMPACT Automates complex research workflows, potentially accelerating scientific discovery by improving efficiency and accuracy in literature analysis.
RANK_REASON The cluster describes a research paper detailing a novel multi-agent system for a specific scientific task.
Read on arXiv cs.IR (Information Retrieval) →
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →