Abstract: The present disclosure relates to a method for identifying a sample data category, an identification model and a method for training it. The identification method comprises: inputting a sample to an identification model obtained by ensembling a plurality of sub-models including a plurality of branch models that are obtained by training full data of a training set through a preset loss function. Each branch model focuses on a focused category subset in the full data. The method further comprises identifying the category to which the input sample belongs according to the output of the identification model. Different loss functions are set through a plurality of branches respectively focusing on different category subsets in the full-scale training set, and the plurality of branches are finally ensembled to obtain an identification model with high identification capability for all data categories of the full data.

Abstract: Described are a system, method, and computer program product for predictive modeling using hyperbolic knowledge graph embeddings. The method includes receiving graph data associated with a knowledge graph including at least one triple. The method also includes generating, in a hyperbolic space, a head embedding for each head vector, a tail embedding for each tail vector, and a relation embedding for each relation vector, of the at least one triple. The method further includes determining a score for each triple based on the head embedding, the tail embedding, and the relation embedding. The method further includes determining a loss based on the score for each triple and updating the head embedding, the tail embedding, and the relation embedding for each triple based on the loss. The method further includes repeating determining the score, determining the loss, and updating until a termination condition is satisfied.

Abstract: A method for automatic optical inspection includes (i) receiving an image of an object, (ii) classifying the image of the object as one of a plurality of categories by a classification model, (iii) determining a label for the image of the object as being qualified if the category obtained by the classification model is a first category, and (iv) performing defect measurement for the image by a segmentation model and determining a label for the image as being qualified or unqualified based on the defect measurement obtained by the segmentation model if the category obtained by the classification model is a second category.

Abstract: Provided is a system that includes a processor to receive a dataset comprising a plurality of feature values of a plurality of features; determine, for each feature of the plurality of features, a plurality of sequence deviation metrics; generate a plurality of sets of features for the plurality of sequence deviation metrics, wherein each set of features comprises a ranked set of features for a sequence deviation metric; train a plurality of machine learning models based on the plurality of sets of features, wherein the plurality of machine learning models comprises a machine learning model for each sequence deviation metric; determine a performance metric of each trained machine learning model for each sequence deviation metric; and select a ranked set of features for a sequence deviation metric that corresponds to a trained machine learning model that has a highest performance metric. Methods and computer program products are also provided.

Abstract: The present disclosure provides a current measurement device. In the device, a first control circuit is connected to a broadband current measurement probe, a quantum current measurement probe includes a magnetic core, a coil provided on the magnetic core, a quantum current measurement member provided in an air gap of the magnetic core, the first control circuit is connected to the coil and the quantum current measurement member respectively; the broadband current measurement probe is configured to obtain a broadband current signal; the first control circuit is configured to determine a coil current value and a regulation microwave frequency according to the broadband current signal, circulate a coil current into the coil according to the coil current value, and control a microwave frequency at which the quantum current measurement member operates according to the regulation microwave frequency; the quantum current measurement member is configured to obtain a measured current value.

Abstract: According to embodiments of the disclosure, a method, an apparatus, a device, and a storage medium for content interaction are provided. The method includes: playing a target video associated with a recommended object; in response to a play progress of the target video reaching a predetermined time point, superimposing and presenting, on the target video, a plurality of images corresponding to a plurality of time points of the target video, the plurality of images being associated with the recommended object; and in response to the play progress of the target video reaching a first time point of the plurality of time points, highlighting a first image corresponding to the first time point in the plurality of images, and non-highlighting other images in the plurality of images except for the first image.

Abstract: Described are a system, method, and computer program product for efficient node embeddings for use in predictive models. The method includes receiving graph data associated with a graph comprising a plurality of nodes associated with a plurality of entities and a plurality of edges associated with interactions between entities. The method also includes generating a plurality of node embeddings for the plurality of nodes, and generating a matrix based on each positive pair of nodes and the plurality of node embeddings. The method further includes decomposing the matrix to provide a left unitary matrix, a diagonal matrix, and a right unitary matrix. The method further includes determining a plurality of updated node embeddings for the plurality of nodes based on the left unitary matrix and the diagonal matrix. The method further includes communicating the plurality of updated node embeddings for inputting into a machine learning model to generate a prediction.

Abstract: A battery separator and a preparation method therefor, and a secondary battery, relating to the technical field of battery separators. The battery separator comprises: a base film; and an inorganic coating provided on at least one surface of the base film. A contact ratio k of inorganic particles to the base film in the longitudinal section of the battery separator is 15-90%; and the relationship among the contact ratio k, air permeability A1 of the base film and air permeability A2 of the coating per unit thickness meets: A1×k?A2. The separator is improved to a certain extent in peel strength, puncture strength and air permeability, and the battery separator with both safety and high ion permeability is provided.

Abstract: A method includes performing a fabrication process that fabricates a wafer having an upper region and unit areas arranged in rows along a first direction and columns along an orthogonal second direction and respective scribe streets between adjacent unit areas to: form first and second electrical components on or in the upper region in respective unit areas or scribe streets, the first and second electrical components spaced apart from one another and including structural features with different respective first and second spacing distances along the first direction.

Abstract: A method for determining a scanning protocol in medical imaging and a medical imaging system are provided. The method includes determining a scanning protocol based on at least two of the following: occlusion information of key points on an examined subject, connection status information between a coil covering the surface of the examined subject and a medical scanning device, and orientation information of the examined subject. The scanning protocol is used by the medical scanning device to scan the examined subject, and the method further includes notifying the scanning protocol. This approach eliminates the need to manually load the scanning protocol, simplifying the scanning process and improving scanning efficiency.

Abstract: Methods, systems, and computer program products for universal depth graph neural networks may obtain a graph G including an adjacency matrix A and an attribute matrix X and train a graph convolutional network using according to the following updating Equation: H=?(?dXW), where H is an output embedding matrix, ?(?) is a nonlinear activation function, ? is a filter, X is an attribute matrix, W is a trainable weight matrix, and d is a trainable depth parameter.

Abstract: A real-time radiometric correction method and an apparatus for spectral reflectance are provided. The method includes: obtaining photosynthetically active irradiance measured by a photosynthetically active radiation sensor, where the photosynthetically active radiation sensor includes a cosine corrector; obtaining photosynthetically active irradiance of a target band based on a proportional relationship between the photosynthetically active irradiance and downwelling solar irradiance within a same wavelength range, where the target band is any band in a spectral band obtained by an imaging hyperspectral sensor, and an imaging time point of the imaging hyperspectral sensor is synchronized with a measurement time point of the photosynthetically active radiation sensor; and performing weighted adjustment on the photosynthetically active irradiance of the target band to obtain corrected target irradiance of the target band.

Abstract: A vacuum switching valve and a suction system having the same. The vacuum switching valve comprises: a valve body, comprising a first end and a second end, the second end being provided with an air inlet, an air outlet and a through hole; a valve element movably arranged in the valve body; a cylinder, the cylinder being connected to the first end and the valve element, the cylinder drives the valve element to move in the valve body, to close or open the air inlet; a stopper passing through the through hole, the stopper comprising a third end and a fourth end, the third end being connected to the valve element, the fourth end being located on the side of the through hole away from the valve element.

Abstract: A system for identifying movers in an independent cart system includes movers having at least one magnet and sensors generating a feedback signal responsive to detecting a magnetic field from the magnet as each mover travels past the sensor. A memory stores an identifier and a corresponding digital fingerprint for each mover. The stored digital fingerprint is generated as a function of the magnetic field generated by the magnet on each mover. A controller receives the feedback signal from each sensor and determines a run-time digital fingerprint for each mover corresponding to the magnetic field generated by the magnet on each mover as a function of the feedback signal. The run-time digital fingerprint is matched to one of the stored digital fingerprints, and the identifier, corresponding to the stored digital fingerprint matching the run-time fingerprint, is read from memory.

Abstract: An image processing method and apparatus may be provided. The process may include, obtaining source compressed texture data of a target image by encoding the target image by using a source compressed texture format and determining a target compressed texture format adapted to the display card and a target compressed block size corresponding to the target compressed texture format. The process may also include obtaining a plurality of pieces of image texture data, based on decoding and aligning the source compressed texture data by using the target compressed block size, and obtaining target compressed texture data of the target image, based on transcoding each piece of the plurality of pieces of image texture data by using the target compressed texture format and the target compressed block size. The process may also include rendering the target image based on the target compressed texture data.

Abstract: Active semiconductor devices in an integrated circuit are provided lateral electrical isolation by surrounding narrow deep trench isolation regions that are merged at shared portions of the narrow deep trench isolation regions. A wide deep trench isolation region laterally surrounds the merged narrow deep trench isolation regions.

Abstract: Provided are computer-implemented methods for generating embeddings for objects which may include receiving heterogeneous network data associated with a plurality of objects in a heterogeneous network; selecting at least one pattern of objects; determining instances of each pattern of objects based on the heterogeneous network data; generating a pattern matrix for each pattern of objects based on the instances of the pattern of objects; generating pattern sequence data associated with a portion of each pattern matrix; generating network sequence data associated with a portion of the heterogeneous network data; and combining the pattern sequence data and the network sequence data into combined sequence data. In some non-limiting embodiments or aspects, methods may include generating a vector for each object of the plurality of objects based on the combined sequence data. Systems and computer program products are also provided.

Abstract: Provided are methods for enhancing a distribution of graph feature embeddings in an embedding space to improve discrimination of graph features by a graph neural network (GNN) that may include receiving a dataset comprising graph data associated with a graph, calculating a distance between a first set of node embeddings and a second set of node embeddings, determining a measure of uniformity for the dataset, determining a plurality of groups of node embeddings, determining a measure of alignment for the plurality of groups of node embeddings, generating a set of graph features based on the measure of uniformity, the measure of alignment, and the distance, and training the GNN based on the set of graph features to provide a trained GNN. Systems and computer program products are also disclosed.

Abstract: A speaker includes a frame, a magnetic circuit system with a magnetic gap, and a vibration system, the vibration system includes a diaphragm assembly, a voice coil assembly, a first elastic wave assembly fixed, and a second elastic wave assembly, the diaphragm assembly includes a dome, a suspension, and a connecting wall extending from the dome to the voice coil assembly, the connecting wall includes a connecting portion connected with the dome the suspension, and a first suspension portion and a second suspension portion located at a side away from the dome, one end of the first elastic wave assembly away from the frame is connected with the first suspension portion, one end of the second elastic wave assembly away from the frame is connected with the second suspension portion. Compared with the related art, the speaker disclosed by the present disclosure could improve space utilization.

Abstract: A developing cartridge includes a casing configured to accommodate developer and having a first side and a second side oppositely arranged in a first direction, and a third side and a fourth side oppositely arranged in the second direction, the first direction and the second direction being intersected with each other. The developing cartridge further includes a developing roller, positioned at the third side, for rotating around a first axis extending in the first direction; a coupling, positioned at the first side, for receiving driving-force output from an image forming device to rotate; a driving surface, inclined to the first direction, including a first driving surface and a second driving surface; a detected part, at least partially positioned at the second side, for moving according to rotation of the coupling.