ComProScanner: A multi-agent based framework for composition-property structured data extraction from scientific literature

Key Points

• The research aims to develop a multi-agent framework for efficient extraction of structured data from scientific literature regarding material compositions and properties.
• Developed a multi-agent based system for data extraction
• Focused on literature related to synthesized materials
• Utilized data-driven techniques to analyze material properties
• Successfully extracted structured data from various scientific sources
• Enhanced accessibility to material compositions and properties data
• Improved the efficiency of data retrieval processes for material scientists