Abstract: A distributor includes: an outer housing; a direct current charging interface, an electronic control terminal interface, and a battery terminal interface disposed in the outer housing. A first contactor and a second contactor are connected between the electronic control terminal interface and the battery terminal interface. A third contactor and a pre-charge resistor form a pre-charge branch. A fourth contactor is connected between the direct current charging interface and the battery terminal interface; a fifth contactor is connected between a negative terminal of the direct current charging interface and a negative terminal of the battery terminal interface; and the five contactors and a pre-charge resistor are disposed in the outer housing.

Abstract: An object is to provide a moving route determination device that makes it possible to accurately determine a user's moving route. A server device 100 includes a geofence evaluation unit 105 configured to store a scoring table for determining a moving route of a user terminal 200, a check-in log acquisition unit 103 configured to acquire at least one of a check-in log which is a history of position management information, such as check-in information, settlement information, or search information, indicating that the user terminal 200 is located in a predetermined range, a settlement log, and a search log as a position management log, and a scoring determination unit 106 configured to determine a moving route of the user terminal 200 on the basis of the position management log and determination information stored in the geofence evaluation unit 105 in a case where a moving means is not able to be determined from at least one of the check-in log, the settlement log, and the search log.

Abstract: The present application discloses a method for acquiring a network resource, device, and system, relating to the field of communications technologies, and enabling logical use of a communication resource while improving the success ratio of communication resource requests. In the present application, a terminal device may request a communication resource from a second network device according to action indication information and/or statistical information sent by a first network device. The terminal device chooses, according to an actual status, to lower a resource request level or to implement other measures matching an actual current status of a network, and then requests a communication resource from the second network device, thereby making logical use of a communication resource while improving the success ratio of communication resource requests.

Abstract: A method, computer system and computer program product for model compression service. The method comprises determining an initial deep neural network (DNN) and an associated compression algorithm available in a compression engine, a type of target hardware and a performance requirement of target hardware. The method also comprises emulating a plurality of different compressed models of the initial DNN on target hardware of the type to obtain corresponding runtime performance data, wherein the different compressed models are defined with different configuration data. The method further comprises obtaining a runtime performance estimator of the target hardware by regression with the different configuration data and the corresponding runtime performance data. Lastly, the method comprises applying the runtime performance estimator to the compression algorithm by the compression engine to generate a compressed DNN of the initial DNN complying with the performance requirement of the type of target hardware.

Abstract: An intelligent wearable product is provided which includes a housing, a power supply circuit, a functional component, and an electrochromic film. A transparent portion is disposed on the housing, and the functional component and the power supply circuit are located inside the housing. The functional component is configured to collect an external parameter by using the transparent portion. The electrochromic film is disposed on an inner surface of the housing and shields the transparent portion. The power supply circuit is configured to supply power to the electrochromic film and the functional component, so that the electrochromic film is transparent when the functional component works. When the intelligent wearable product is not worn on a human body, the electrochromic film and the housing jointly present an integrated visual effect. According to the application, quality, competitiveness and appearance experience of the intelligent wearable product are improved.

Abstract: Methods and systems for training a machine learning model include training a machine learning model using training data. A status of the machine learning model's training is determined based on an accuracy curve of the machine learning model over the course of the training. Parameters of the training are adjusted based on the status. Training of the machine learning model is completed using the adjusted parameters.

Abstract: Using a training portion of a dataset, a set of component parameters comprising parameters of a component of an object detection model are trained. Using the trained set of component parameters, a set of backbone component weights comprising weights of component types in a backbone portion of the object detection model are trained. Using the trained set of component parameters, a set of backbone link weights comprising weights of links within the backbone portion are trained. Using the trained set of component parameters, a set of head component weights comprising weights of component types in a head portion of the object detection model are trained. Using the trained sets of component parameters, backbone component weights, backbone link weights, and head component weights, a trained object detection model is configured and trained to perform object detection.

Abstract: A model parallel training technique for neural architecture search including the following operations: (i) receiving a plurality of ML (machine learning) models that can be substantially interchangeably applied to a computing task; (ii) for each given ML model of the plurality of ML models: (a) determining how the given ML model should be split for model parallel processing operations, and (b) computing a model parallelism score (MPS) for the given ML model, with the MPS being based on an assumption that the split for the given ML model will be used at runtime; and (iii) selecting a selected ML model based, at least in part, on the MPS scores of the ML models of the plurality of ML models.

Abstract: Multi-level objectives improve efficiency of multi-objective automated machine learning. A hyperband framework is established with a kernel density estimator to shrink the search space based on evaluation of lower-level objectives. A Gaussian prior assumption directly shrinks the search space to find a main objective.

Abstract: A one-shot neural architecture search method referred to as MergeNAS by merging different types of convolutions into a single operation. This mergence approach reduces the search cost to roughly half a GPU-day as well as alleviates the over-fitting problem caused by a traditional differentiable architecture search (DARTS) approach by reducing the number of redundant parameters.

Abstract: A method for optimizing a neural network architecture by estimating an inference time for each operator in the neural network architecture is provided. The method may include determining a benchmark time for at least one single-path architecture out of a plurality of single-path architectures associated with the neural network by sampling the at least one single-path architecture from the neural network, wherein the at least one single-path architecture comprises one or more operators. The method may further include, based on the benchmark time for the at least one single-path architecture, determining an estimated inference time for an operator, wherein determining the estimated inference time for the operator comprises, applying an operator function, wherein the operator function comprises a function based on a difference between the benchmark time associated with the at least one single-path architecture and the estimated latency of the neural network.

Abstract: Methods, computer program products, and/or systems are provided that perform the following operations: obtaining a target neural network structure and constraints for a target neural network; generating a meta learning network having an associated quantization function based, at least in part, on the target neural network structure; training the meta learning network based, at least in part, on providing a hybrid quantization vector as input to the meta learning network and providing a training dataset to the target neural network; obtaining a plurality of hybrid quantization vectors; determining a new hybrid quantization vector from the plurality of hybrid quantization vectors; and retraining the trained meta learning network based, at least in part, on providing the new hybrid quantization vector as input to the trained meta learning network.

Abstract: Aspects of the invention include methods and systems that include obtaining a source domain dataset. The source domain dataset includes corresponding labels, and the source domain dataset and the corresponding labels are associated with training a source domain machine learning model. A method includes obtaining a target domain dataset without corresponding labels and a feature vector that identifies features in the source domain dataset and the target domain dataset. The method also includes obtaining a set of loss terms from known machine learning models that implement a domain adversarial neural network (DANN) architecture. The DANN architecture includes feed-forward propagation and backpropagation. A target domain machine learning model is obtained based on the source domain dataset, the target domain dataset, the feature vector, and the set of loss terms and without labels for the target domain dataset to perform training.

Abstract: The disclosure provides a multi-object tracking algorithm based on an object detection and feature extraction combination model, including the following steps: S1, adding an object appearance feature extraction network layer behind a prediction feature layer of an object detection tracking network having an FPN structure; S2, calculating object fused loss of the object detection tracking network having the FPN structure and added with the object appearance feature extraction network layer; S3, forming a feature comparison database utilizing a neural network during multi-frame objection detection and tracking process; and S4, comparing current image object appearance features with features in the feature comparison database, drawing an object trajectory if the objects are uniform; else adding the current image object appearance features into the feature comparison database to form a new feature comparison database, and then repeating steps S2 and S3.

Abstract: The disclosure provides a feature compression algorithm based on neural network, including the following steps: S1, image data preparation: collecting facial images, and uniformly performing map processing to the facial images collected; S2, feature data acquisition: delivering the facial images processed into a face recognition system for face detection and feature extraction, and saving facial feature data; S3, setting up a neural network model; S4, model iteration training; S5, storing a parameter model; and S6, feature compression. The feature compression algorithm based on neural network of the disclosure can not only achieve compression of original feature data, but also retain its original semantic feature, which belongs to a higher-dimensional feature abstraction. The compressed feature data can be directly used.

Abstract: Aspects of the invention include systems and methods to obtain meta features of a dataset for training in a deep learning application. A method includes selecting an initial search space that defines a type of deep learning architecture representation that specifies hyperparameters for two or more neural network architectures. The method also includes applying a search strategy to the initial search space. One of the two or more neural network architectures are selected based on a result of an evaluation according to the search strategy. A new search space is generated with new hyperparameters using an evolutionary algorithm and a mutation type that defines one or more changes in the hyperparameters specified by the initial search space, and, based on the mutation type, the new hyperparameters are applied to the one of the two or more neural networks or the search strategy is applied to the new search space.

Abstract: An object is to provide a moving route determination device that makes it possible to accurately determine a user's moving route. A server device 100 includes a geofence evaluation unit 105 configured to store a scoring table for determining a moving route of a user terminal 200, a check-in log acquisition unit 103 configured to acquire at least one of a check-in log which is a history of position management information, such as check-in information, settlement information, or search information, indicating that the user terminal 200 is located in a predetermined range, a settlement log, and a search log as a position management log, and a scoring determination unit 106 configured to determine a moving route of the user terminal 200 on the basis of the position management log and determination information stored in the geofence evaluation unit 105 in a case where a moving means is not able to be determined from at least one of the check-in log, the settlement log, and the search log.