YOUR GO-TO INTELLIGENCE RESOURCE FOR THE SCHOLARLY PUBLISHING INDUSTRY

Elsevier and LG AI Research announced that researchers can now identify chemical information previously limited to images, drawings, and reaction schemes in patents and scientific literature.

Visual chemistry has historically been difficult to index at scale. The chemistry‑specific AI vision technology developed by LG AI Research is now integrated into Elsevier’s content extraction and curation processes for Reaxys, the company’s discovery chemistry solution. This enables faster and more accurate capture of substance information from images in patent and journal content, improving the scale and precision of Elsevier’s curation of substances, reactions, bioactivities, biological targets, and substance properties.

Much of the chemical data scientists rely on is communicated through figures and reaction schemes rather than searchable text. When this information is not searchable, researchers often need to manually review documents to confirm whether a compound or reaction has been described. Making such chemistry discoverable at scale is a complex AI challenge, as chemical drawings encode meaning through bonds, atoms, stereochemistry, and spatial relationships. Misinterpretation of these elements can lead to incomplete or inaccurate results, particularly in areas such as novelty searching, competitive intelligence, synthesis planning, and inorganic or organometallic chemistry.

The AI vision technology from LG AI Research combines molecule detection, reaction‑diagram parsing, and optical chemical structure recognition (OCSR) in a single model. Benchmarking results published by the company indicate that it outperforms alternatives in extracting chemistry from full document pages. Each extraction pipeline is validated against Reaxys benchmarks before deployment to ensure accuracy is maintained alongside scalability. The pipeline has undergone extensive testing across Elsevier’s data and workflow tools prior to wider implementation.

The collaboration will next extend to reaction extraction, broadening image‑based data capture beyond individual substances. Elsevier and LG AI Research are also exploring additional challenges to address together, combining LG AI Research’s AI expertise with Elsevier’s chemistry content and curation capabilities.

This initiative aligns with Elsevier’s Responsible AI Principles and Privacy Principles.