AutoSynthesis: An agentic system for automated meta-analysis
AutoSynthesis takes a natural-language research question and runs the meta-analysis pipeline through to a PRISMA-style report.
The paper describes a multi-agent system that searches literature, screens studies, checks eligibility, extracts quantitative statistics, computes standardized effect sizes, and performs random-effects meta-analysis. It also supports moderator-based heterogeneity analysis and risk-of-bias assessment. In the authors’ application, it screened more than 28 studies and extracted over 20 quantitative claims, with pooled estimates close to expert-conducted meta-analyses using Hedges’ g. ArXiv · AI/CL/LG's note
The paper describes a multi-agent system that searches literature, screens studies, checks eligibility, extracts quantitative statistics, computes standardized effect sizes, and performs random-effects meta-analysis. It also supports moderator-based heterogeneity analysis and risk-of-bias assessment. In the authors’ application, it screened more than 28 studies and extracted over 20 quantitative claims, with pooled estimates close to expert-conducted meta-analyses using Hedges’ g. ArXiv · AI/CL/LG's note
score 5