Artificial intelligence evaluates chemical spectra in minutes
Artificial intelligence evaluates chemical spectra in minutes
Publish Date: 2026-06-23 00:16:00
Source Domain: www.chemeurope.com
- Determining the chemical structure of substances, especially novel compounds, is a time-consuming task for chemists.
- Chemists use analytical techniques like NMR spectroscopy, infrared spectroscopy, and mass spectrometry to deduce structures, often solving complex “chemical puzzles” from signals.
- Novel molecules pose special challenges as measurement data can be imperfect and may include interferences from impurities.
- The new SECS AI system combines two AI approaches to translate spectra into a mathematical representation and refines molecular structures based on measurement data.
- SECS produces a ranked list of possible structures with similarity scores, performing at a level comparable to human specialists in identifying molecular structures.
- SECS is an adjunct tool complementing human expertise rather than replacing it, offering a second opinion on structure elucidation.
- The tool is openly accessible, with the source code, model data, and a test version available for research purposes.
