The article from Strieth-Kalthoff et al. explores the integration of artificial intelligence (AI) with expert knowledge in the field of retrosynthetic planning, which is crucial for designing complex chemical syntheses. The authors argue that while AI has made significant strides in various scientific fields, its application in chemical synthesis planning has been limited by the quality and quantity of available reaction data. They emphasize that a hybrid approach, combining AI's computational power with the nuanced understanding of human chemists, is essential for overcoming these limitations. This synergy can lead to more accurate and efficient synthesis planning by leveraging the strengths of both data-driven algorithms and expert insights.

Synthia^® exemplifies this hybrid approach by utilizing a database of expert-coded chemical synthesis rules alongside AI algorithms. This combination allows Synthia^® to generate robust and viable synthetic pathways more effectively than relying on machine learning alone. By integrating expert knowledge into its algorithms, Synthia^® can provide chemists with innovative and efficient routes for chemical synthesis, supporting both drug discovery and process optimization.

Read the full article


Artificial Intelligence for Retrosynthetic Planning Needs Both Data and Expert Knowledge
Felix Strieth-Kalthoff, Sara Szymkuć, Karol Molga, Alán Aspuru-Guzik, Frank Glorius, and Bartosz A. Grzybowski
Journal of the American Chemical Society 2024 146 (16), 11005-11017
DOI: 10.1021/jacs.4c00338