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: Systems and methods are provided for increasing penetration rates of connected vehicles in traffic regions of interest. An example method may comprise: (1) determining a target connected vehicle penetration rate for a traffic region to implement a traffic management strategy; (2) determining trips for connected vehicles to achieve the target connected vehicle penetration rate for the traffic region, wherein the determined trips are predicted to bring the connected vehicles to the traffic region within a target time interval; and (3) autonomously controlling the connected vehicles in accordance with the determined trips. Here, determining a respective trip for a respective connected vehicle may comprises at least one of: )a) changing an original departure time for a pre-planned trip of the respective connected vehicle; or (b) changing an original navigation route for the pre-planned trip.

Abstract: A preparation method of a sulfur-containing biphenylsulfur-containing biphenyl compound shown by the general formula (I) has a reaction formula as follows: Each substituent in the formula is defined in the description. In the method, halogeno benzene shown by the general formula (II) generates a coupling reaction in a catalytic system composed of a nickel compound and at least one ligand under the combined action of metallic zinc to obtain the sulfur-containing biphenyl compound shown by the general formula (I).

Abstract: Devices with metal structures formed with seams and methods of fabrication are provided. An exemplary method includes forming a metal plug having a top surface formed with a seam; depositing a film over the top surface of the metal plug and at least partially filling the seam; and etching the film from over the metal plug, wherein the film remains in the seam.

Abstract: Aspects of the presently disclosed technology may be implemented to provide systems and methods for incentive-based vehicle routing. For example, a method may include determining a suggested route for a vehicle based on a location of a congested area of a road network to decrease congestion of the road network. The method may include determining an individualized incentive for an occupant of the vehicle to incentivize the occupant to follow the suggested route. The method may include transmitting, to the vehicle or to a device of the occupant, a message indicating the suggested route and the individualized incentive for following the suggested route, with the message providing an option for the occupant to accept the suggested route. The method may include, upon the occupant accepting the suggested route, updating an original route of the vehicle to the suggested route thereby directing the vehicle along the suggested route.

Abstract: The present application relates to the technical field of power grid ice-melting, in particular to an AC short-circuit ice-melting control method and an ice-melting distribution transformer. The present application includes a distribution transformer with dual functions of voltage reduction and ice-melting, an ice-melting control system and its control method. The transformer is equipped with two groups of low-voltage output terminals, which are respectively used for power distribution supply and ice-melting operations; the ice-melting control system includes a local intelligent control box and a remote communication control box, and uses a deep reinforcement learning model to calculate the optimal ice-melting power gear and ice-melting time.

Abstract: Dimension reduction is performed for multimodal embedding searches. A request to perform a search using input data of a modality of a multimodal machine learning model is received. The input data is encoded using the multimodal machine learning model to generate a first embedding that represents the input data. A machine learning dimension reduction technique is applied to the first embedding to generate a second embedding with a reduced number of dimensions. This second embedding is then used to search an index of data objects also generated using both the multimodal machine learning model and the machine learning dimension reduction technique.

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.