Abstract: Systems, apparatuses and methods may provide for conducting an importance measurement of a plurality of parameters in a trained neural network and setting a subset of the plurality of parameters to zero based on the importance measurement. Additionally, the pruned neural network may be re-trained. In one example, conducting the importance measurement includes comparing two or more parameter values that contain covariance matrix information.

Abstract: Methods and apparatus for discrimitive semantic transfer and physics-inspired optimization in deep learning are disclosed. A computation training method for a convolutional neural network (CNN) includes receiving a sequence of training images in the CNN of a first stage to describe objects of a cluttered scene as a semantic segmentation mask. The semantic segmentation mask is received in a semantic segmentation network of a second stage to produce semantic features. Using weights from the first stage as feature extractors and weights from the second stage as classifiers, edges of the cluttered scene are identified using the semantic features.

Abstract: Provided is an LED support. The LED support includes a substrate, an inner electrode, a heat dissipation pad, an electrode pad, and a solder resist layer. The electrode pad includes a top connection layer and a bottom electrode layer that are laminated. The top connection layer is connected to the back of the substrate and is provided with a connection portion extending toward the heat dissipation pad and connected to the conductive through hole in the substrate. The connection portion is between the heat dissipation pad and the bottom electrode layer. The spacing between the bottom electrode layer and the heat dissipation pad is greater than the spacing between the connection portion and the heat dissipation pad. The solder resist layer is filled in a groove formed between the electrode pad and the heat dissipation pad on the back of the substrate, and covers the connection portion.

Abstract: A lead frame for an optocoupler is provided, which is a metal sheet and includes multiple frame units and multiple mutually parallel connecting strips. The frame units are arranged at intervals on each of the connecting strips. Each of the frame units includes two adjacent functional parts, and two pins perpendicularly connected to the connecting strip. The two functional parts are both located on one side of the connecting strip and are respectively connected to the two pins and ends of the two pins are located on the other side of the connecting strip. The area of each of the functional parts is sufficient to accommodate a chip to be installed and fixed, and the chip is any one of a light-emitting chip and a light-sensing chip.

Abstract: Methods and systems for budgeted and simplified training of deep neural networks (DNNs) are disclosed. In one example, a trainer is to train a DNN using a plurality of training sub-images derived from a down-sampled training image. A tester is to test the trained DNN using a plurality of testing sub-images derived from a down-sampled testing image. In another example, in a recurrent deep Q-network (RDQN) having a local attention mechanism located between a convolutional neural network (CNN) and a long-short time memory (LSTM), a plurality of feature maps are generated by the CNN from an input image. Hard-attention is applied by the local attention mechanism to the generated plurality of feature maps by selecting a subset of the generated feature maps. Soft attention is applied by the local attention mechanism to the selected subset of generated feature maps by providing weights to the selected subset of generated feature maps in obtaining weighted feature maps.

Abstract: The present application discloses a few-shot image recognition method and apparatus, a device, and a storage medium. The method includes: obtaining to-be-recognized images, and constructing an image episode according to the to-be-recognized image, the image episode including a support set and a query set; inputting the image episode into a pre-trained image recognition model, the image recognition model being a few-shot image recognition model based on hard episode training; and calculating a similarity between an image in the query set and each class in the support set according to the image recognition model, and determining the class of to-be-recognized images in the query set according to the similarity.

Abstract: A high-efficiency supercritical oil extraction method is provided. According to the method, in an early extraction stage, extraction is performed using an extractant along an extraction kettle from top to bottom; in a middle extraction stage, the extraction kettle is longitudinally flipped by 180 degrees, extraction is still performed using an extractant along the extraction kettle from top to bottom; in a later extraction stage, extraction is performed using a conventional extractant along the extraction kettle from bottom to top. Thus, the pressure effect and the critical extraction effect of the critical extractant are simultaneously exerted, thereby greatly improving the extraction efficiency of supercritical oil extraction, shortening the extraction time, and reducing the dosage of extractant, so as to achieve high-efficiency and low-cost oil extraction.

Abstract: A high-efficiency supercritical oil extraction method is provided. According to the method, in an early extraction stage, extraction is performed using an extractant along an extraction kettle from top to bottom; in a middle extraction stage, the extraction kettle is longitudinally flipped by 180 degrees, extraction is still performed using an extractant along the extraction kettle from top to bottom; in a later extraction stage, extraction is performed using a conventional extractant along the extraction kettle from bottom to top. Thus, the pressure effect and the critical extraction effect of the critical extractant are simultaneously exerted, thereby greatly improving the extraction efficiency of supercritical oil extraction, shortening the extraction time, and reducing the dosage of extractant, so as to achieve high-efficiency and low-cost oil extraction.