Abstract: The present disclosure discloses a detection method and system of power equipment based on a multispectral image, and relates to the field of smart grid information technology. The method includes: obtaining an image of power equipment to be detected, where the image is one of an infrared image, an ultraviolet image, and a visible image; inputting the image into a pre-trained pixel-based power equipment detection model for detection, and performing classified prediction on pixels in the image to obtain a predicted result; and outputting a predicted image based on the predicted result, where the predicted image is power equipment image with background information removed, and is marked with a name of each piece of equipment. With the implementation of the present disclosure, efficiency and accuracy of power equipment detection can be improved.

Abstract: A forecasting system is designed to train and utilize dynamic models for forecasting metrics associated with content provision. For example, a forecasting system may identify characteristics of received training data relating to a particular title. The characteristics may be used to select one of a plurality of available forecasting models to train.

Abstract: A sensor and housing assembly may include a sensor configured to detect a quantity of a gas in engine exhaust, and a housing including an inner upstream wall, a first inner side wall having a sensor opening for the sensor, an inner downstream wall opposite to the inner upstream wall, and a second inner side wall opposite to the first inner side wall, and having a housing inlet and a housing outlet that is opposite to the sensor opening in the first inner side wall. The sensor is positioned within the sensor opening on the first inner side wall, and a tip of the sensor protrudes through the housing outlet.

Abstract: A LiDAR device including a emitting unit configured to emit a laser, a detecting unit configured to detect the laser, and a controller configured to control the emitting unit and the detecting unit, wherein the controller is configured to operate the LiDAR device at in at least a first operation mode or a second operation mode in which power consumption per unit time is lower than that of the first operation mode, and wherein the controller is configured to switch an operation mode of the LiDAR device from the first operation mode to the second operation mode when the controller receives a first trigger signal from a first device, and switch the operation mode of the LiDAR device from the second operation mode to the first operation mode when the controller receives a second trigger signal from the first device.

Abstract: The present specification discloses an image diagnosis apparatus using a deep learning model and a method therefor, wherein tissue included in an input medical image is classified as being one of normal and abnormal for a disease by using the trained deep learning model using a weighted loss function in which different weights are assigned to a probability distribution of determining that a picture extracted from the input medical image is abnormal even though it is normal and a probability distribution of determining that the picture is normal even though it is abnormal.

Abstract: Proposed is a method for processing, by one or more processors, data obtained on the basis of detection signals generated by a detector array including a plurality of detector units. The method includes generating, on the basis of the detection signals generated by the detector array, a spatio-temporal data set including a plurality of counting values, wherein each of the plurality of counting values corresponds to one of the plurality of detector units and is addressed to at least one time bin, and generating distance information including a plurality of distance values corresponding to each of the plurality of detector units by processing the spatio-temporal data set.

Abstract: Systems, apparatuses, methods, and computer program products are disclosed for unsupervised named entity recognition. An example method includes receiving, by a communications circuitry, a reference named entity list, the reference named entity list identifying a set of named entities and an entity type of each identified named entity. The example method further includes generating, by a vectorizer, vectors from the named entities identified in the reference named entity list, and consolidating, by a synthesizer, the generated vectors into a set of representative vectors, wherein each representative vector is associated with a particular entity type. Finally, the example method receiving, by an analysis engine, a set of text, and performing, by the analysis engine, named entity recognition on the set of text using the set of representative vectors to generate a tagged set of text.

Abstract: A resonance conversion device and a universal serial bus circuit are provided. The resonance conversion device includes an input filter circuit, a full-bridge LLC converter circuit, a transformer circuit, a rectifier filter circuit, and a controller. The controller is configured to determine whether an indication voltage of a voltage command is less than or equal to a threshold value and to perform the following steps: in response to the indication voltage being less than or equal to the threshold value, controlling the full-bridge LLC converter circuit into a half-bridge operation mode, and regulating the DC output voltage by performing half-bridge burst mode control on the full-bridge LLC converter circuit based on the DC output voltage; and in response to the indication voltage being greater than the threshold value, regulating the DC output voltage by performing full-bridge burst mode control on the full-bridge LLC converter circuit based on the DC output voltage.

Abstract: In general, this disclosure describes techniques for generating and evaluating automatic transcripts of audio recordings containing human speech. In some examples, a computing system is configured to: generate transcripts of a plurality of audio recordings; determine an error rate for each transcript by comparing the transcript to a reference transcript of the audio recording; receive, for each transcript, a subjective ranking selected from a plurality of subjective rank categories; determine, based on the error rates and subjective rankings, objective rank categories defined by error-rate ranges; and assign an objective ranking to a new machine-generated transcript of a new audio recording, based on the objective rank categories and an error rate of the new machine-generated transcript.

Abstract: A planar winding transformer includes a magnetic core set and a multilayer circuit board. The magnetic core set includes two magnetic cores and two magnetic columns. The two magnetic cores are parallel to each other. The multilayer circuit board is disposed between two magnetic cores, and two magnetic columns penetrate through the multilayer circuit board. The multilayer circuit board includes two low voltage winding layers and one high voltage winding layer. Two low voltage winding layers are connected to each other in parallel, and the high voltage winding layer is disposed between two low voltage winding layers. When the high voltage winding layer receives a polarity current, at least one of the low voltage winding layers generates a corresponding induced current. Two magnetic cores and two magnetic columns form a closed path for magnetic flux.

Abstract: Techniques are presented for optimizing a treatment plan for charged particle therapy. The method includes obtaining medical image data voxels inside a subject in a reference frame of a radiation source that emits a beam of charged particles at multiple tracks with a controlled emitted energy at each track. Hydrogen density (HD) is determined based on the medical image data. Stopping power ratio (SPR) along a first beam having a first track and first emitted energy is calculated based on HD. A range to a Bragg peak is calculated along the first beam based on the SPR and the first emitted energy. The first beam track or the first emitted energy, or both, is modified based at least in part on the beam range to determine a second track and second emitted energy. Output data that indicates the second track and second emitted energy are sent.

Abstract: A method for processing point data obtained from a light detection and ranging (LiDAR) device is proposed. The method may include obtaining point cloud data including a plurality of point data for a plurality of detection points. The method may also include generating an image of the plurality of detection points based on the point cloud data. Each of the plurality of point data may include location information about a detection point and a geometrically enhanced intensity of the detection point. The geometrically enhanced intensity may be generated based on a combination of a reflection parameter related to the amount of light scattered at the detection point and a geometric parameter based on geometrical characteristic of the detection point.

Abstract: In general, this disclosure describes techniques for generating and evaluating automatic transcripts of audio recordings containing human speech. In some examples, a computing system is configured to: generate transcripts of a plurality of audio recordings; determine an error rate for each transcript by comparing the transcript to a reference transcript of the audio recording; receive, for each transcript, a subjective ranking selected from a plurality of subjective rank categories; determine, based on the error rates and subjective rankings, objective rank categories defined by error-rate ranges; and assign an objective ranking to a new machine-generated transcript of a new audio recording, based on the objective rank categories and an error rate of the new machine-generated transcript.

Abstract: An engine exhaust aftertreatment system is disclosed. The system may comprise: an internal combustion engine having an intake passage and an exhaust passage; a turbocharger fluidly connected to the internal combustion engine, the turbocharger including a compressor and a turbine, the compressor being in fluid communication with the intake passage, and the turbine being in fluid communication with the exhaust passage; a reductant injector situated downstream of the turbocharger, wherein the reductant injector is closely coupled to the turbocharger such that a reductant is injected into an exhaust flow of the turbocharger; a first container downstream of the reductant injector, the first container including a multi-functional catalyst (MFC); and a second container downstream of the first container, the second container including a selective catalytic reduction (SCR) component and an Ammonia catalyst (AMOx) component.

Abstract: Sharing sensor data between a first device and a second device comprising obtaining a set of point data from at least one of sensors located in the first device, determining a property data of the subset of point data based on the subset of point data, generating a first sharing data for sharing with the second device based on the property data, transmitting the sharing data to the second device, identifying an occurrence of an event at a first time point, and generating a second sharing data different from the first sharing data. A content of the second sharing data includes at least a portion of the set of point data obtained within a first time period including the first time point.

Abstract: Sharing sensor data of a first device with a second device includes obtaining a set of point data from at least one of a sensors located in the first device, generating a first property data of the first subset of point data based on the first subset of point data, generating a sharing data including at least a portion of the first subset of point data and the first property data, and transmitting the sharing data to the second device. If a class of a first object included in the class information a class in which personal information must be protected, a content of the sharing data includes a privacy protection data in which the first subset of point data is processed such that personal information of the first object does not identified by the second device.

Abstract: An engine exhaust aftertreatment system is disclosed. The system may comprise: an internal combustion engine having an intake passage and an exhaust passage; a turbocharger fluidly connected to the internal combustion engine, the turbocharger including a compressor and a turbine, the compressor being in fluid communication with the intake passage, and the turbine being in fluid communication with the exhaust passage; a reductant injector situated downstream of the turbocharger, wherein the reductant injector is closely coupled to the turbocharger such that a reductant is injected into an exhaust flow of the turbocharger; a first container downstream of the reductant injector, the first container including a multi-functional catalyst (MFC); and a second container downstream of the first container, the second container including a selective catalytic reduction (SCR) component and an Ammonia catalyst (AMOx) component.

Abstract: Systems and methods for detecting demographic bias in automatic speech recognition (ASR) systems. Corpuses of transcriptions from different demographic groups are analyzed, where one of the groups is known to be susceptible to bias and another group is known not to be susceptible to bias. A difference between the transcription accuracy for the first group and a transcription accuracy for a second group is measured. ASR accuracy for each group is measured and compared to each other using both statistics-based and practicality-based methodologies to determine whether a given ASR system or model exhibits a meaningful level of bias. Based on the statistical significance and the practical significance, an alert including a recommendation to adjust the ASR model is generated.

Abstract: A method for determining an initial rotor position of a permanent magnet synchronous motor (PMSM) includes: generating a plurality of transient currents by applying a plurality of voltages to each phase stator winding of a three phase stator winding of the PMSM; generating three phase current differences according to the plurality of transient currents; determining a first zone in which the initial rotor position of the PMSM is located according to the three phase current differences, wherein angles between 0-360 degrees are divided into a plurality of zones, and the first zone is selected from the plurality of zones; calculating three line current differences according to the three phase current differences; and determining the initial rotor position of the PMSM according to the first zone and the three line current differences.