Details

Autor(en) / Beteiligte

• Strieth‐Kalthoff, Felix
• Sandfort, Frederik
• Kühnemund, Marius
• Schäfer, Felix R.
• Kuchen, Herbert
• Glorius, Frank

Titel

Machine Learning for Chemical Reactivity: The Importance of Failed Experiments

Ist Teil von

• Angewandte Chemie International Edition, 2022-07, Vol.61 (29), p.e202204647-n/a

Auflage

International ed. in English

Ort / Verlag

Germany: Wiley Subscription Services, Inc

Erscheinungsjahr

2022

Quelle

Alma/SFX Local Collection

Beschreibungen/Notizen

• Assessing the outcomes of chemical reactions in a quantitative fashion has been a cornerstone across all synthetic disciplines. Classically approached through empirical optimization, data‐driven modelling bears an enormous potential to streamline this process. However, such predictive models require significant quantities of high‐quality data, the availability of which is limited: Main reasons for this include experimental errors and, importantly, human biases regarding experiment selection and result reporting. In a series of case studies, we investigate the impact of these biases for drawing general conclusions from chemical reaction data, revealing the utmost importance of “negative” examples. Eventually, case studies into data expansion approaches showcase directions to circumvent these limitations—and demonstrate perspectives towards a long‐term data quality enhancement in chemistry. Learning from failure? The performance of machine learning models depends critically on the data quality. This study investigates the possibility of learning chemical reactivity from reaction databases. In model experiments, the impact of experimental errors, biased experiment selection and incomplete result reporting are analyzed for predicting cross‐coupling yields. These results showcase directions—and what it takes to enhance data quality.