Abstract: A charging system includes at least one first charging device and at least one second charging device that are cascaded. The first charging device includes at least one first accommodation portion for mounting a first energy storage device and a connecting port for connecting the second charging device or the first charging device. The second charging device includes at least one second accommodation portion for mounting a second energy storage device and a connecting port for connecting the first charging device or the second charging device. The first accommodation portion and the second accommodation portion have different shapes or dimensions.

Abstract: A charging system includes at least one first charging device and at least one second charging device that are cascaded. The first charging device includes at least one first accommodation portion for mounting a first energy storage device and a connecting port for connecting the second charging device or the first charging device. The second charging device includes at least one second accommodation portion for mounting a second energy storage device and a connecting port for connecting the first charging device or the second charging device. The first accommodation portion and the second accommodation portion have different shapes or dimensions.

Abstract: The present disclosure relates to a method for recognizing facial expressions based on adversarial elimination. First, a facial expression recognition network is built based on a deep convolutional neural network. On a natural facial expression data set, the facial expression recognition network is trained through a loss function to make facial expression features easier to distinguish. Then some key features of input images are actively eliminated by using an improved confrontation elimination method to generate a new data set to train new networks with different weight distributions and feature extraction capabilities, forcing the network to perform expression classification discrimination based on more features, which reduces the influence of interference factors such as occlusion on the network recognition accuracy rate, and improving the robustness of the facial expression recognition network.

Abstract: A seesaw-type feeding device for collecting solid materials suspended in a flow field includes a suspension assembly, a material collection assembly, a joint hinge, two anti-collision blocks and an escapement shaking assembly, wherein the latter three of which are located between the suspension assembly and the material collection assembly. The suspension assembly includes a V-shaped floating body and at least three fixed bottom anchors. The material collection assembly includes a collection pocket net, a material container, a delivery part, and an annular water discharging screen mesh. The escapement shaking assembly is located behind the V-shaped floating body and faces the material collection assembly. The material collection assembly is able to shake relatively to the fluid in the flow field, and the escapement shaking assembly drives the material collection assembly to shake.

Abstract: A method for multi-spectral scattering-matrix tomography includes a step of splitting an input light signal into an incident light signal and a reference light signal. The sample light signal is directed to a sample in either a reflection configuration or a transmission configuration such that an output light signal includes light scattered from or transmitted through the sample. The incident signal and the reference light signal are directed to a camera angled to allow for amplitude and phase to be calculated by off-axis holography. A total light signal is measured with a camera that is a coherent sum of the reference light signal and the output signal. The total light signal for each light frequency and each incident angle are collected as collected total light signal data. A computing device derives an image of the sample from a calculated reflection matrix or transmission matrix or both of them.

Abstract: The present disclosure discloses an adaptive high-precision compression method and system based on a convolutional neural network model, and belongs to the fields of artificial intelligence, computer vision, and image processing. According to the method of the present disclosure, coarse-grained pruning is performed on a neural network model by using a differential evolution algorithm first, and the coarse-grained space is quickly searched through an entropy importance criterion and an objective function with good guidance to obtain a near-optimal neural network structure. Then fine-grained search space is built on the basis of an optimal individual obtained from the coarse-grained search, and fine-grained pruning is performed on the neural network model by a differential evolution algorithm to obtain a network model with an optimal structure. Finally, the performance of the optimal model is restored by using a multi-teacher multi-step knowledge distillation network to reach the precision of an original model.

Abstract: The present disclosure relates to a method for recognizing facial expressions based on adversarial elimination. First, a facial expression recognition network is built based on a deep convolutional neural network. On a natural facial expression data set, the facial expression recognition network is trained through a loss function to make facial expression features easier to distinguish. Then some key features of input images are actively eliminated by using an improved confrontation elimination method to generate a new data set to train new networks with different weight distributions and feature extraction capabilities, forcing the network to perform expression classification discrimination based on more features, which reduces the influence of interference factors such as occlusion on the network recognition accuracy rate, and improving the robustness of the facial expression recognition network.

Abstract: A distributed sender driven Admission Control System (ACS) is described herein, leveraging Weighted-Fair Quality of Service (QoS) queues, found in standard NICs and switches, to guarantee RPC level latency service level objectives (SLOs) by a judicious selection of QoS weights and traffic-mix across QoS queues. ACS installs cluster-wide RPC latency SLOs by mapping LS RPCs to higher weight QoS queues, and coping with overloads by adaptively apportioning LS RPCs amongst QoS queues based on measured completion times for each queue. When the network demand spikes unexpectedly to predetermined threshold percentage of provisioned capacity, ACS achieves a latency SLO that is significantly lower than the state-of-art congestion control at the 99.9th-p and admits significantly more RPCs meeting SLO target when RPC sizes are not aligned with priorities.

Abstract: A cyber security system employing machine learning techniques to help predict and protect computing systems from cyber-attacks. The system comprises data sources for storing security data, a deployment infrastructure for generating a portion of the security data insights, and a data analytics module for processing the security data. The data analytics module includes a data connector unit for collecting and organizing the security data into a selected format, a data preprocessing unit for cleaning the organized security data, a cyber feature unit for identifying based on preselected cyber features selected portions of the cleaned security data associated with the cyber features, a model development unit for applying one or more selected machine learning techniques to the features to form output model data, and a model prediction unit for generating based on the output model data one or more prediction values based on the cleaned security data and the cyber features.

Abstract: The present technology addresses a need in the art for an automated and scalable mechanism to authorize a containerized process. An aspect of the present technology deals with authorizing an unprivileged process by a privileged process without embedding credentials or network access at the time of validation. The present technology provides the possibility for the privileged process to continuously (dynamically) validate the authenticity of the unprivileged process by performing a plurality of operations to ensure the unprivileged process has maintained its authenticity while having access to sensitive information.

Abstract: Embodiments of the disclosure disclose a control method and apparatus for displaying a widget, a system, an electronic device, a server and a storage medium. The method includes: acquiring a displaying control parameter of a widget on a page, generating content of the widget based on a state parameter of a displaying state in response to detecting a current time reaches an effective time of the displaying state, and displaying the widget whose state is updated to the displaying state based on the content of the widget.

Abstract: A cloud-end data multicast method includes: obtaining a multicast packet, the multicast packet carrying a tenant identifier, a destination address, and a source address; and searching for a multicast group according to the tenant identifier and the destination address. The multicast group includes multicast members. The method also includes obtaining routes corresponding to the multicast members; generating a routing tree according to the routes corresponding to multicast members; and obtaining member addresses corresponding to the multicast members. The method also includes obtaining an address list containing addresses of target members that are identified to receive the multicast packet by performing address filtering according to the source address and the member addresses; encapsulating the multicast packet; and delivering the encapsulated multicast packet to the target members according to the address list and the routing tree.

Abstract: Disclosed in the present invention is an application of soluble protein BAFF in B cell in-vitro culture and proliferation. Peptide expressed by secretory BAFF peptide in the present invention can be effectively secreted out of cells to generate corresponding bioactivity. During the process of culture, a cell line in the present invention can continuously secrete water-soluble BAFF protein having bioactivity, the water-soluble BAFF protein served as tool cells are co-cultured with B cells, and in-vitro proliferation and antigen presentation effects of human B cells can be effectively improved.

Abstract: Embodiments provide a system including a first host computing device that includes a first virtual machine (VM) and a first application. The system also includes a second host computing device including a virtualization software layer, a second VM, and an auto-discovery service at least partially instantiated within the virtualization software layer. The auto-discovery service is configured to receive a message and an auto-discovery packet from a second application executing on the second VM. The auto-discovery service inserts an option into the auto-discovery packet, and transmits the auto-discovery packet to the first application. The option in the auto-discovery packet includes the message received from the second application.

Abstract: The present invention relates to a system and method for improving glucose sensor accuracy by utilizing multiple calibration methods and selecting the most accurate method depending on a consensus glucose concentration estimate. Embodiments of the present invention comprise the steps of performing at least one in vivo update of surrounding glucose to acquire glucose values; calculating multiple updated calibration estimates using the updated glucose values; calculating an initial consensus glucose estimate from sensor output using each updated calibration estimate; applying a smooth crossover function to the multiple calibration estimates based on the value of the initial consensus glucose estimate; and adding weights to the multiple calibration estimates to acquire a consensus glucose estimate.

Abstract: A cloud-end data multicast method includes: obtaining a multicast packet, the multicast packet carrying a tenant identifier, a destination address, and a source address; and searching for a multicast group according to the tenant identifier and the destination address. The multicast group includes multicast members. The method also includes obtaining routes corresponding to the multicast members; generating a routing tree according to the routes corresponding to multicast members; and obtaining member addresses corresponding to the multicast members. The method also includes obtaining an address list containing addresses of target members that are identified to receive the multicast packet by performing address filtering according to the source address and the member addresses; encapsulating the multicast packet; and delivering the encapsulated multicast packet to the target members according to the address list and the routing tree.

Abstract: An ink for forming a compound semiconductor thin film is provided, which contains a binder includes a compound includes an S atom or an Se atom and metallic compound particles which are both dispersed in an organic solvent. A compound semiconductor thin film is formed by applying or printing the ink for forming a compound semiconductor thin film and heat-treating it. A solar cell is constituted, which has a light-absorbing layer formed of the compound semiconductor thin film.

Abstract: A method includes the steps of performing a coating or printing of ink for producing a compound semiconductor thin film so as to form a compound semiconductor coating film, the ink including 50% by mass or more of amorphous compound nanoparticles, mechanically applying a pressure to the compound semiconductor coating film, and subjecting the compound semiconductor coating film to a heat-treatment to form a compound semiconductor thin film.

Abstract: A method for producing a nanoparticle for forming a CZTS compound semiconductor thin film is provided which includes the step of reacting a solution including a metal salt or a metal complex with a solution including a chalcogenide salt to produce a CZTS compound nanoparticle. A CZTS compound semiconductor thin film is formed by coating or printing the nanoparticle for forming the CZTS compound semiconductor thin film, and subjecting it to a heat treatment. A solar cell including the CZTS compound semiconductor thin film as the light-absorbing layer is provided.

Abstract: The present invention relates to a system and method for improving glucose sensor accuracy by utilizing multiple calibration methods and selecting the most accurate method depending on a consensus glucose concentration estimate. Embodiments of the present invention comprise the steps of performing at least one in vivo update of surrounding glucose to acquire glucose values; calculating multiple updated calibration estimates using the updated glucose values; calculating an initial consensus glucose estimate from sensor output using each updated calibration estimate; applying a smooth crossover function to the multiple calibration estimates based on the value of the initial consensus glucose estimate; and adding weights to the multiple calibration estimates to acquire a consensus glucose estimate.