Abstract: A method for crystal pulling is provided. The method includes the following. The method includes performing an equal-diameter process. The equal-diameter process is performed as follows. From the first equal-diameter stage to the third equal-diameter stage, the crystal rotation rate is gradually increased after starting at a first initial crystal rotation rate, and then is kept at a constant rotation rate after gradually increasing the crystal rotation rate. The crucible rotation rate is gradually increased from a first initial crucible rotation rate to a maximum crucible rotation rate after starting at the first initial crucible rotation rate in the first equal-diameter stage, the crucible rotation rate is kept at the maximum crucible rotation rate in the second equal-diameter stage, and the crucible rotation rate is gradually decreased after the second equal-diameter stage. The method further includes a cooling process.

Abstract: The single crystal furnace charging system includes a control system configured to send a charging request, a material preparation system configured to receive the charging request and prepare materials based on the charging request, a feeding system configured to receive the charging request, obtain the materials, and compare an actual feeding amount with a preset feeding amount to obtain a difference between the actual feeding amount and the preset charging amount, in response to the difference being within a first preset threshold, charge a single crystal furnace. The single crystal furnace charging system further includes a calibration system configured to obtain a total charging weight, and obtain a difference between the total charging weight and a preset charging amount of the single crystal furnace, in response to the difference being within a second preset threshold, update a total charging amount to the total charging weight.

Abstract: An adaptively adjusted and accurate parking control method for an ATO system includes performing statistical learning on the basis of historical stop information, and adaptively inferring a parking point offset. Moreover, two application preconditions of the method are provided: firstly, speed tracking performance is good in an electric braking stage, and secondly, a pneumatic braking process has random and statistical stationary characteristics. The method improves the average stop accuracy in a statistical sense, satisfies a high-accuracy stop requirement of the entire train formation, and can also evaluate train performances and track circumstances timely and duly, thereby satisfying complex and variable real-time operation task requirements.

Abstract: Various embodiments are described herein for a system for analyzing images and speech obtained during a medical diagnostic procedure for automatically generated annotated images using annotation data for one or more images 5 having at least one object of interest (OOI) and a classification where the annotation data includes text that is generated from speech provided by the user commenting on the one or more images having the at least one OOI.

Abstract: Technology for inspection for detecting a defect of a printed matter using machine logic informed by machine learning. Some embodiments of the present invention may include one, or more, of the following features: (i) generates defect datasets; (ii) generates defect libraries; (iii) uses the generated defect libraries for deep learning training; and (iv) uses machine learning to detect defects using computer code (for example, a *.jpg format file) corresponding to an image of a piece of printed matter instead of using a visual image (that is, an image of the type that is created when a person takes a picture using a traditional film camera).

Abstract: A method for crystal pulling is provided. The method includes the following. The method includes performing an equal-diameter process. The equal-diameter process is performed as follows. From the first equal-diameter stage to the third equal-diameter stage, the crystal rotation rate is gradually increased after starting at a first initial crystal rotation rate, and then is kept at a constant rotation rate after gradually increasing the crystal rotation rate. The crucible rotation rate is gradually increased from a first initial crucible rotation rate to a maximum crucible rotation rate after starting at the first initial crucible rotation rate in the first equal-diameter stage, the crucible rotation rate is kept at the maximum crucible rotation rate in the second equal-diameter stage, and the crucible rotation rate is gradually decreased after the second equal-diameter stage. The method further includes a cooling process.

Abstract: Parameters from an inspection image representing mechanical vibrations and electromagnetic interference can be determined. An X-direction vibration spectrum can be determined based on the X-direction offsets. A Y-direction vibration spectrum can be determined based on the Y-direction offsets. The determinations can be based on a swath image of a workpiece, such as a semiconductor wafer or reticle.

Abstract: The single crystal furnace charging system includes a control system configured to send a charging request, a material preparation system configured to receive the charging request and prepare materials based on the charging request, a feeding system configured to receive the charging request, obtain the materials, and compare an actual feeding amount with a preset feeding amount to obtain a difference between the actual feeding amount and the preset charging amount, in response to the difference being within a first preset threshold, charge a single crystal furnace. The single crystal furnace charging system further includes a calibration system configured to obtain a total charging weight, and obtain a difference between the total charging weight and a preset charging amount of the single crystal furnace, in response to the difference being within a second preset threshold, update a total charging amount to the total charging weight.

Abstract: A method, system and computer program product for segmenting generic foreground objects in images and videos. For segmenting generic foreground objects in videos, an appearance stream of an image in a video frame is processed using a first deep neural network. Furthermore, a motion stream of an optical flow image in the video frame is processed using a second deep neural network. The appearance and motion streams are then joined to combine complementary appearance and motion information to perform segmentation of generic objects in the video frame. Generic foreground objects are segmented in images by training a convolutional deep neural network to estimate a likelihood that a pixel in an image belongs to a foreground object. After receiving the image, the likelihood that the pixel in the image is part of the foreground object as opposed to background is then determined using the trained convolutional deep neural network.

Abstract: Embodiments of this application provide a communication method, an apparatus, and a system. The method includes: A proxy base station receives a MSG1 from a user equipment. The proxy base station sends a MSG2 to the user equipment in response to the MSG1, where the MSG2 includes a timing advance, and the timing advance is determined based on at least one of an internal processing delay of a signal extension system and a delay between the signal extension system and a signal source base station.

Abstract: Technology for inspection for detecting a defect of a printed matter using machine logic informed by machine learning. Some embodiments of the present invention may include one, or more, of the following features: (i) generates defect datasets; (ii) generates defect libraries; (iii) uses the generated defect libraries for deep learning training; and (iv) uses machine learning to detect defects using computer code (for example, a *.jpg format file) corresponding to an image of a piece of printed matter instead of using a visual image (that is, an image of the type that is created when a person takes a picture using a traditional film camera).

Abstract: A transport protein coding gene, and a method for efficient production of L-tryptophan by a strain containing the gene. Specifically, by heterologous expression of ywkB gene from Bacillus subtilis on the genome of Escherichia coli, L-tryptophan production efficiency of the strain can be improved. Performing shake flask fermentation with the strain can accumulate 15.2 g/L of L-tryptophan within 24 h, which is 35% higher than a control strain.

Abstract: Disclosed are gene engineering bacteria for producing L-arginine and a construction method and an application of the gene engineering bacteria. According to the method, genes encoding a carbamoyl phosphate synthetase and a gene encoding an L-arginine biosynthesis pathway enzyme are integrated into Escherichia coli; the present invention has analyzed and reconstructed the arginine synthetic pathway and the metabolic flow related to arginine in the entire amino acid metabolic network in E. coli and finally obtained a genetically engineered bacterial strain which has a clear genetic background, carries no plasmids, undergoes no mutagenesis and is capable of stably and efficiently producing L-arginine.

Abstract: A method, system and computer program product for segmenting generic foreground objects in images and videos. For segmenting generic foreground objects in videos, an appearance stream of an image in a video frame is processed using a first deep neural network. Furthermore, a motion stream of an optical flow image in the video frame is processed using a second deep neural network. The appearance and motion streams are then joined to combine complementary appearance and motion information to perform segmentation of generic objects in the video frame. Generic foreground objects are segmented in images by training a convolutional deep neural network to estimate a likelihood that a pixel in an image belongs to a foreground object. After receiving the image, the likelihood that the pixel in the image is part of the foreground object as opposed to background is then determined using the trained convolutional deep neural network.

Abstract: A method, system and computer program product for segmenting generic foreground objects in images and videos. For segmenting generic foreground objects in videos, an appearance stream of an image in a video frame is processed using a first deep neural network. Furthermore, a motion stream of an optical flow image in the video frame is processed using a second deep neural network. The appearance and motion streams are then joined to combine complementary appearance and motion information to perform segmentation of generic objects in the video frame. Generic foreground objects are segmented in images by training a convolutional deep neural network to estimate a likelihood that a pixel in an image belongs to a foreground object. After receiving the image, the likelihood that the pixel in the image is part of the foreground object as opposed to background is then determined using the trained convolutional deep neural network.