Turbidity measurements are influenced by environmental factors such as water temperature. We designed experiments to study whether water temperature affects in-line turbidity detection and the potential influence mechanism. A turbidity meter installed in-line could self-record data, including the water temperature and turbidity values. From our experimental analysis, we verified the influence of water temperature on the in-line turbidity. Moreover, the temperature coefficient should not be obtained from the experiment directly because the intrinsic impact of in-line turbidity detection does not come from water temperature. Instead, the effect is derived from the optical components’ heat change. When the water temperature change is insignificant, the in-line turbidity deviation caused by the water temperature can be ignored. However, when the water temperature changes substantially, the in-line turbidity sensor should compensate for the temperature drift.

Predicting products of organic chemical reactions is useful in chemical sciences, especially when one or more reactants are new organics. However, the performance of traditional learning models heavily relies on high-quality labeled data. In this work, to utilize unlabeled data for better prediction performance, we propose a method that combines semi-supervised learning with graph convolutional neural networks for chemical reaction prediction. First, we propose a Mean Teacher Weisfeiler–Lehman Network to find the reaction centers. Then, we construct the candidate product set. Finally, we use an Improved Weisfeiler–Lehman Difference Network to rank candidate products. Experimental results demonstrate that, with 400k labeled data, our framework can improve the top-5 accuracy by 0.7% using 35k unlabeled data. When the proportion of unlabeled data increases, the performance gain can be larger. For example, with 80k labeled data and 35k unlabeled data, the performance gain with our framework can be 1.8%.