Automatic detection of steel surface defects is very important for product quality control in the steel industry. However, the traditional method cannot be well applied in the production line, because of its low accuracy and slow running speed. The current, popular algorithm (based on deep learning) also has the problem of low accuracy, and there is still a lot of room for improvement. This paper proposes a method combining improved ResNet50 and enhanced faster region convolutional neural networks (faster R-CNN) to reduce the average running time and improve the accuracy. Firstly, the image input into the improved ResNet50 model, which add the deformable revolution network (DCN) and improved cutout to classify the sample with defects and without defects. If the probability of having a defect is less than 0.3, the algorithm directly outputs the sample without defects. Otherwise, the samples are further input into the improved faster R-CNN, which adds spatial pyramid pooling (SPP), enhanced feature pyramid networks (FPN), and matrix NMS. The final output is the location and classification of the defect in the sample or without defect in the sample. By analyzing the data set obtained in the real factory environment, the accuracy of this method can reach 98.2%. At the same time, the average running time is faster than other models.

Given motorized spindles’ extensive periods of prolonged high-velocity operation, they are prone to temperature changes, which leads to the problem of thermal error, leading to diminished precision in machining operations. To address the thermal error issue in motorized spindles of computer numerical control (CNC) machine tools, this study proposes a pelican optimization algorithm (POA)-optimized convolutional neural network (CNN)–long short-term memory (LSTM) hybrid neural network model (POA-CNN-LSTMNN). Initially, the identification of temperature-sensitive locations in the spindle system is performed using a combination of hierarchical clustering, the K-medoids algorithm, and Pearson’s coefficient calculation. Subsequently, the temperature data from these identified points, along with real-time collected spindle thermal error data, are employed to construct the model. The Pelican optimization algorithm is used to enhance the model parameters to achieve the best performance. Finally, the proposed model is subjected to a comparative analysis with other thermal error prediction models. Drawing from the experimental findings, it is evident that the POA-CNN-LSTMNN model exhibits superior prediction performance.

To better solve the problem of thermal error of computerized numerical control machining equipment (CNCME), a thermal error prediction model based on the sparrow search algorithm and long short-term memory neural network (SSA-LSTMNN) is proposed. Firstly, the Fuzzy C-means clustering algorithm (FCMCA) is used to screen the key temperature-sensitive points of the CNCME. Secondly, by taking the temperature rise data of key temperature-sensitive points as input and the corresponding time thermal error data as output, we established the SSA-LSTMNN thermal error prediction model. The SSA is used to optimize the parameters of LSTMNN and make its performance play the best. Taking the VMC1060 vertical machining center as the research object, we carried out the experiment. Finally, the prediction effect of the proposed model is compared with the article swarm optimized algorithm and LSTM neural network (PSOA-LSTMNN), the LSTMNN, and the traditional recurrent neural network (TRNN) model. The results show that the average values of the predicted residual fluctuations of the SSA-LSTMNN model are all more than 44% lower than those of the other three models under different operating conditions, which has a strong practicality.

Extracting relation quintuple and feature words from unstructured text is a prelude to the construction of the scientific knowledge base. At present, the prior works use explicit clues between entities to study this task but ignore the use and the association of the feature words. In this work, we propose a new method to generate self-adaptive feature words from the original text for every single sample. These words can add additional correlation information to the knowledge graph. We allow the model to generate a new word representation and apply it to the original sentence to judge the relation type and locate the head and tail of the relation quintuple. Compared with the previous works, the feature words increase the flexibility of relying on information and improve the explanatory ability. Extensive experiments on scientific field datasets illustrate that the self-adaptive feature words method (SAFW) is good at ferreting out the unique feature words and obtaining the core part for the quintuple. It achieves good performance on four public datasets and obtains a markable performance improvement compared with other baselines.