Four labs find experimental differences can undermine AI catalyst predictions
To turn abundant carbon dioxide into valuable fuel, we need a fast and efficient way to determine which catalysts work best over the longest time. AI models have the potential to help guide catalyst selection, but as with internet chatbots, AI models are only as good as the data you put into them.
Four laboratories collaborated to test a rhodium-based catalyst in hydrogenation experiments, aiming to improve AI models for catalyst selection. However, their efforts resulted in inconsistent yields of carbon monoxide and methane, complicating the data for machine-learning models. The discrepancy arose largely due to differences in mixing intensity during the experiments.
Despite the challenges, the researchers identified standardized reactor design, protocols, and operating conditions as crucial for enhancing experimental reproducibility. These findings emphasize the importance of meticulous experimental design when feeding data into AI models, as variability can significantly impact the reliability of predictions.
Written by urgent.news from Phys.org's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
