Under the low-carbon background, with the aid of the Malmquist–Luenberger SBM (Slack-based Measure) model of unexpected output, the green total factor productivity (GTFP) of the logistics industry in Jiangsu Province, China, was measured and decomposed in this study based on the reality and experience of logistics industry development in 13 cities in three regions of Jiangsu Province in the years 2006–2018 by taking resource consumption into the input system and discharged pollutants into the output system. It is concluded that the environmental regulation (ER) has a significant positive effect on the growth of the GTFP of the logistics industry, and technological progress has become an important endogenous force that promotes the GTFP of the logistics industry in Jiangsu Province. On this basis, a dynamic GMM (Generalized method of moment) model and a Tobit model were constructed to further study the possible temporal and spatial effects of ER on the GTFP of the logistics industry. The research results reveal that the ER can exert both promoting and inhibitory effects on the GTFP of the logistics industry, and there is a temporal turning point for the effects. Besides, the effects notably differ spatially and temporally. Finally, some policies and advice for the green sustainable development of the logistics industry were proposed. For example, the government and enterprises should pay attention to the green and efficient development of the logistics industry and dynamically adjust the ER methods. They should consider the greening of both forward logistics links and reverse logistics system in the supply chain.

Nonintrusive load monitoring (NILM) analyzes only the main circuit load information with an algorithm to decompose the load, which is an important way to help reduce energy usage. Recent research shows that deep learning has become popular for this problem. However, the ability of a neural network to extract load features depends on its structure. Therefore, more research is required to determine the best network architecture. This study proposed two deep neural networks based on the attention mechanism to improve the current sequence to point (s2p) learning model. The first model employs Bahdanau style attention and RNN layers, and the second model replaces the RNN layer with a self-attention layer. The two models are both based on a time embedding layer. Therefore, they can be better applied in NILM. To verify the effectiveness of the algorithms, we selected two open datasets and compared them with the original s2p model. The results show that attention mechanisms can effectively improve the model’s performance.