Abstract: A method for characterizing polarization image information and a method for computing characterization parameters are provided. The method includes: obtaining n polarized subimages of a polarization imaging target, wherein each polarized subimage corresponds to a different polarization angle and n?3; and computing a parameter matrix [ILP (x, y) INLP(x, y) ?(x, y)] of a polarization cosine characterization equation of all pixel points of the polarization imaging target according to all polarized subimages and the polarization cosine characterization equation of the polarization imaging target. The disclosure obtains a maximum polarization intensity image and a minimum polarization intensity image through computing multiple polarization intensity images, which improves the accuracy of polarization imaging, and the degree of accuracy is much higher than macroscopic accuracy.

Abstract: Provided in the present application are a data processing method and apparatus based on data coding, and a device. The method comprises: performing N-channel error correction coding on data to be processed that is in information to be processed, so as to obtain N pieces of coded data to be processed; by using coded meta-channel data obtained by means of error correction coding, performing redundancy processing on the coded data to be processed, so as to obtain N pieces of response data; and then performing error correction decoding on the N pieces of response data, so as to obtain processing result information of the information to be processed.

Abstract: A method for measuring drop time of a control rod cluster integrated with a rod position measurement device is provided, wherein the method is used to measure the drop time of each control rod cluster, and includes: Si, monitoring a voltage Ua of coils in Group A to capture a rod cluster drop signal; S2, searching a point (tmax, Vmax) with a maximum drop speed or with a local maximum drop speed; S3, retroactively calculating, from tmax, an end of a time period T4 when the control rod cluster starts to drop; S4, retroactively searching, from a minimum value point of a drop reference signal DROPref, a start of the time period T4 when the drop reference signal DROPref drops from a maximum value to 33% thereof; and S5, determining, from tmax forward, a time point t6 when a drop speed of the control rod cluster is lower than 0.

Abstract: All example composition may include a plurality of grains including an iron nitride phase. The plurality of grains may have an average wain size between about 10 nm and about 200 nm. An example technique may include treating a composition including a plurality of grains including au iron-based phase to adjust an average grain size of the plurality of grains to between about 20 nm and about 100 ma. The example technique may include nitriding the plurality of grains to form or grow an iron nitride phase.

Abstract: The disclosure relates to the technical field of vehicles, and particularly provides a brake caliper assembly, a brake, and a vehicle, aiming to solve the problem of how to increase the maximum deflection that a leaf spring can bear on the premise of ensuring an elastic restoring force of the leaf spring. The brake caliper assembly of the disclosure comprises a leaf spring arranged between a first brake caliper and a second brake caliper. The leaf spring comprises an elastic structure. The elastic structure comprises a bending portion, an extending portion, an outward-expanding portion, and an abutting portion that are connected in sequence; the outward-expanding portion extends to the outside of the leaf spring from the extending portion, the bending portion is bent inwards, and the abutting portion abuts on the first brake caliper and/or the second brake caliper.

Abstract: The disclosed embodiments provide a system for processing data. During operation, the system applies a first set of hash functions to a first entity identifier (ID) for a first entity to generate a first set of hash values. Next, the system produces a first set of intermediate vectors from the first set of hash values and a first set of lookup tables by matching each hash value in the first set of hash values to an entry in a corresponding lookup table in the first set of lookup tables. The system then performs an element-wise aggregation of the first set of intermediate vectors to produce a first embedding. Finally, the system outputs the first embedding for use by a machine learning model.

Abstract: The disclosed embodiments provide a system for processing data. During operation, the system obtains a training dataset containing a first set of records associated with a first set of identifier (ID) values and an evaluation dataset containing a second set of records associated with a second set of ID values. Next, the system selects a random subset of ID values from the second set of ID values. The system then generates a sampled evaluation dataset comprising a first subset of records associated with the random subset of ID values in the second set of records. The system also generates a sampled training dataset comprising a second subset of records associated with the random subset of ID values in the first set of records. Finally, the system outputs the sampled training dataset and the sampled evaluation dataset for use in training and evaluating a machine learning model.

Abstract: The disclosure relates to the technical field of vehicle accessories, and in particular to a brake for a vehicle and a vehicle. The disclosure aims at solving the problems present in existing brakes for a vehicle that oil passages are relatively long, an angle at which each oil passage is arranged is relatively large, debris generated in a machining process remains in the oil passages and then causes blockage of the oil passages, and drilling holes in an inner caliper body and an outer caliper body to form the oil passages further reduces mechanical strength of the inner caliper body and the outer caliper body.

Abstract: The disclosure relates to the technical field of vehicle accessories, and particularly provides a brake for a vehicle and a vehicle. The disclosure aims at solving the problem that existing brakes for a vehicle usually cannot overcome low-frequency stubborn noise generated by a caliper structure in the brake in a braking process.

Abstract: Techniques for training and using a neural network to make predictions using trajectory modelling are disclosed herein. In some embodiments, a computer-implemented method comprises: training a first neural network with a first machine learning algorithm using training data, the first neural network being a recurrent neural network, the training data including a plurality of reference career trajectories, each reference career trajectory in the plurality of reference career trajectories comprising a sequence of reference career segments, each reference career segment in the sequence of reference career segments comprising reference profile data and reference time data indicating a position of the reference career segment within the sequence of reference career segments, the training data also including a corresponding set of reference skills for each reference career segment.

Abstract: In an example embodiment, a warm-start training solution is used to dramatically reduce the computational resources needed to train when retraining a generalized additive mixed-effect (GAME) model. The problem of retraining time is particularly applicable to GAME models, since these models take much longer to train as the data grows. In the past, the strategy to reduce computational resources during retraining was to use less training data, but this affects the model quality, especially for GAME models, which rely on fine-grained sub-models at, for example, member or item levels. The present solution addresses the computational resources issues without sacrificing GAME model accuracy.

Abstract: The disclosure provides a low-nickel nitrogen-containing austenitic stainless steel flux-cored wire and a preparation method thereof, and belongs to the technical field of welding materials. The disclosure aims at solving the technical problems of nitrogen element loss, air holes, hot cracks in a welding seam area, and pitting corrosion caused by nitride precipitation in a heat affected area that are easily generated in a welding joint when low-nickel nitrogen-containing austenitic stainless steel is welded in the prior art. The flux-cored wire of the disclosure is prepared from a flux core and a stainless steel sheath. During welding, gas protection is not needed. The flux core is formed by mixing an alloy component and a slag system. The alloy component is formed by mixing electrolytic manganese, ferrosilicon, chromium metal, nickel metal, ferromolybdenum, copper powder and ferrochrome nitride powder in percentage by mass.

Abstract: This invention presents methods for MU-MIMO wireless communication systems comprising a BS with plural of antennas, either closely located or distributed; A plural of AFRs deployed over a coverage area, each AFR has NBF?1 BF antennas with a beam pattern facing the MU-BFer and NUF?1 UF antennas with a beam pattern facing UEs or downstream AFR(s); A channel estimation module for estimating the Total Channels between a plural of UEs and the BS with the AFRs in place; and, A MU-BFer that uses the estimates of the Total Channels to perform beamforming computations for transmitting and/or receiving multiple spatially multiplexed streams of signals to or from a plural of UEs using the same frequency resource.

Abstract: The disclosed embodiments provide a system for processing data. During operation, the system obtains a training dataset containing a first set of records associated with a first set of identifier (ID) values and an evaluation dataset containing a second set of records associated with a second set of ID values. Next, the system selects a random subset of ID values from the second set of ID values. The system then generates a sampled evaluation dataset comprising a first subset of records associated with the random subset of ID values in the second set of records. The system also generates a sampled training dataset comprising a second subset of records associated with the random subset of ID values in the first set of records. Finally, the system outputs the sampled training dataset and the sampled evaluation dataset for use in training and evaluating a machine learning model.

Abstract: This invention presents methods for MU-MIMO wireless communication systems comprising a BS with plural of antennas, either closely located or distributed; A plural of AFRs deployed over a coverage area, each AFR has N_BF?1 BF antennas with a beam pattern facing the MU-BFer and N_UF?1 UF antennas with a beam pattern facing UEs or downstream AFR(s); A channel estimation module for estimating the Total Channels between a plural of UEs and the BS with the AFRs in place; and, A MU-BFer that uses the estimates of the Total Channels to perform beamforming computations for transmitting and/or receiving multiple spatially multiplexed streams of signals to or from a plural of UEs using the same frequency resource.

Abstract: Herein are techniques for configuring, integrating, and operating trainable tensor transformers that each encapsulate an ensemble of trainable machine learning (ML) models. In an embodiment, a computer-implemented trainable tensor transformer uses underlying ML models and additional mechanisms to assemble and convert data tensors as needed to generate output records based on input records and inferencing. The transformer processes each input record as follows. Input tensors of the input record are converted into converted tensors. Each converted tensor represents a respective feature of many features that are capable of being processed by the underlying trainable models. The trainable models are applied to respective subsets of converted tensors to generate an inference for the input record. The inference is converted into a prediction tensor. The prediction tensor and input tensors are stored as output tensors of a respective output record for the input record.

Abstract: An intelligent cooking machine, comprising a top cover device, a base and a main control device, in which the top cover device is located above the base and is provided with a main material feeding device, a complementary material feeding device, a sauce feeding device, a cooking oil feeding device and a dressing feeding device, the main material feeding device is provided with a weighing sensor, and the weighing sensor is connected with the main control device to generate weight information of main materials contained in the main material feeding device and to allow the main control device to comprehensively control various feeding devices. Arrangement of various feeding devices of the invention is more scientific, comprehensive and reasonable, material feeding is more intelligent and controllable, the cooking effect is better, and the automation, integration and intelligence levels are higher.

Abstract: The disclosed embodiments provide a system for processing data. During operation, the system applies a first set of hash functions to a first entity identifier (ID) for a first entity to generate a first set of hash values. Next, the system produces a first set of intermediate vectors from the first set of hash values and a first set of lookup tables by matching each hash value in the first set of hash values to an entry in a corresponding lookup table in the first set of lookup tables. The system then performs an element-wise aggregation of the first set of intermediate vectors to produce a first embedding. Finally, the system outputs the first embedding for use by a machine learning model.

Abstract: In an example embodiment, a warm-start training solution is used to dramatically reduce the computational resources needed to train when retraining a generalized additive mixed-effect (GAME) model. The problem of retraining time is particularly applicable to GAME models, since these models take much longer to train as the data grows. In the past, the strategy to reduce computational resources during retraining was to use less training data, but this affects the model quality, especially for GAME models, which rely on fine-grained sub-models at, for example, member or item levels. The present solution addresses the computational resources issues without sacrificing GAME model accuracy.

Abstract: An intelligent cooking machine, comprising a top cover device, a base and a main control device, in which the top cover device is located above the base and is provided with a main material feeding device, a complementary material feeding device, a sauce feeding device, a cooking oil feeding device and a dressing feeding device, the main material feeding device is provided with a weighing sensor, and the weighing sensor is connected with the main control device to generate weight information of main materials contained in the main material feeding device and to allow the main control device to comprehensively control various feeding devices. Arrangement of various feeding devices of the invention is more scientific, comprehensive and reasonable, material feeding is more intelligent and controllable, the cooking effect is better, and the automation, integration and intelligence levels are higher.