Abstract: Some embodiments of the disclosure provide a method for detecting an anti-chimeric antigen receptor (CAR) antibody against humanized CAR T cells. In some examples, the method include following steps. S1. Transfecting a humanized CAR into Chinese hamster ovary (CHO) cells into using lentivirus to generate CHO cells expressing CAR as target cells. S2. Collecting samples to be tested. S3. Preparing different concentrations of standards using human proteins corresponding to a target of the CAR and establishing a standard curve. S4. Incubating the target cells with the different concentrations of standards and the samples to be tested, and capturing the anti-CAR antibody. S5. Incubating a phycoerythrin (PE)-conjugated anti-human IgG Fc antibody with CHO cells for labeling. S6. Calculating the anti-CAR antibody by detecting average fluorescence intensities of the standards, blank control, and the samples to be tested using flow cytometry.

Abstract: The present disclosure relates to analysis of Hg(II) isotopes in the aqueous phase, and provides an in-situ enrichment method and an analytical method for Hg(II) isotopes in the aqueous phase. The in-situ enrichment method includes the following steps: conducting adsorption in the water sample by a diffusive gradient in thin-films (DGT) device to obtain an Hg(II)-adsorbed DGT device; where a binding gel of the DGT device is an NSBA gel; conducting elution on the NSBA gel in the Hg(II)-adsorbed DGT device to acquire an Hg(II)-containing eluate with a mercury concentration higher than or equal to 0.5 ng/mL; where the elution is carried out with an eluent of reverse aqua regia. In the in-situ enrichment method, the NSBA-DGT device only needs to be placed in the water sample to be investigated to perform in-situ Hg(II) adsorption without direct grab sampling.

Abstract: A catheter reinforcement layer (100) and a catheter (10). The catheter reinforcement layer (100) includes a spring component (110) and at least one axial component (120). The at least one axial component (120) extends along the spring component (110) from a proximal end to a distal end, and has at least one intersection with the spring component (110). According to the catheter reinforcing layer (100), by introducing one or more axial components (120) into a reinforcing layer of an existing spring component (110), the axial modulus of the catheter (10) is increased, thereby avoiding axial deformation or even breakage of the tubular body of the catheter (10) due to axial force.

Abstract: Systems and methods which provide for tool-free detachable implement change without aid of separate tools are described. A tool-free implement change mechanism may comprise a plurality of apparatuses operable in cooperation to facilitate installation and/or removal of detachable implements with respect to a host garden machine without aid of separate tools. For example, a tool-free implement change mechanism may comprise a motor shaft locking apparatus and an implement securing apparatus. In operation according to embodiments, a motor shaft locking apparatus of embodiments may be engaged to prevent movement of motor shaft when tightening or loosening an implement securing apparatus used for affixing the detachable implement to the shaft assembly.

Abstract: Disclosed are a neural network training method, a neural network training device and an electronic device. The neural network training method includes: training a first neural network to be trained by using sample data; determining an indicator parameter of the first neural network in a current training process; determining an update manner corresponding to a preset condition if the indicator parameter meets the preset condition; and updating a parameter of a batch normalization layer in the first neural network based on the update manner. In this way, sparsing of a feature map output by a neural network is implemented, thereby reducing an amount of data to be transmitted and improving computation speed of a chip.

Abstract: Disclosed is a method for convolution calculation in a neural network, comprising: reading an input feature map, depthwise convolution kernels and pointwise convolution kernels from a dynamitic random access memory (DRAM); performing depthwise convolution calculations and pointwise convolution calculations according to the input feature map, the depthwise convolution kernels and the pointwise convolution kernels to obtain output feature values of a first predetermined number p of points on all pointwise convolution output channels; storing the output feature values of a first predetermined number p of points on all pointwise convolution output channels into an on-chip memory, wherein the first predetermined number p is determined according to at least one of available space in the on-chip memory, a number of the depthwise convolution calculation units, and width, height and channel dimensions of the input feature map; and repeating the above operation obtain output feature values of all points on all pointwis

Abstract: Disclosed is an apparatus for performing a convolution operation in a convolutional neural network. The apparatus may comprise a selector for selecting one or more nonzero elements of a weight parameter, a selector for selecting a data item(s) corresponding to selected nonzero elements in input feature data, and a calculator unit for performing an operation. The apparatus may realize the convolution operation in a sparsified convolutional neural network efficiently through the hardware.

Abstract: A communication method includes receiving, from an indoor unit, an indoor unit model, an indoor unit protocol type, and a priority number; determining a communication protocol type and a communication address corresponding to communication with the indoor unit according to the indoor unit model, the indoor unit protocol type, and the priority number; and sending the communication protocol type and the communication address to the indoor unit for the indoor unit to establish a communicative connection to an outdoor unit according to the communication protocol type and the communication address.

Abstract: Embodiments of the present invention are directed to a computer-implemented method for generating a framework for analyzing adverse drug reactions. A non-limiting example of the computer-implemented method includes receiving to a processor, a plurality of drug chemical structures. The non-limiting example also includes receiving, to the processor, a plurality of known drug-adverse drug reaction associations. The non-limiting example also includes constructing, by the processor, a deep learning framework for each of a plurality of adverse drug reactions based at least in part upon the plurality of drug chemical structures and the plurality of known adverse-drug reaction associations.

Abstract: Embodiments of the present invention are directed to a computer-implemented method for generating a framework for analyzing adverse drug reactions. A non-limiting example of the computer-implemented method includes receiving to a processor, a plurality of drug chemical structures. The non-limiting example also includes receiving, to the processor, a plurality of known drug-adverse drug reaction associations. The non-limiting example also includes constructing, by the processor, a deep learning framework for each of a plurality of adverse drug reactions based at least in part upon the plurality of drug chemical structures and the plurality of known adverse-drug reaction associations.

Abstract: A household appliance networking method is provided. Pursuant to the method, in a household appliance networking mode, a wireless access point name and a password input by a user are obtained. Whether the wireless access point name and the password are correct is verified. Based on a verification that the wireless access point name and the password are correct, a configuration key value input by the user is obtained. Based on a determination that a networking verification according to the configuration key value fails, a target configuration key value in a preset mode is determined and the same is inputted, so as to enable successful household appliance networking. An apparatus, an electronic device, and a storage medium embodying the method are also provided.

Abstract: A method and apparatus for pushing subtitle data in a live scenario. The method includes: obtaining video stream data and audio stream data, the audio stream data being data corresponding to an audio part in the video stream data; generating the subtitle data according to the audio stream data, the subtitle data comprising a subtitle text corresponding to a speech in the audio stream data and time information of the subtitle text; and pushing, in response to pushing the video stream data to a user terminal, the subtitle data to the user terminal, the subtitle data instructing the user terminal to synchronously display the subtitle text with live pictures in the video stream data and the audio part in the audio stream data according to the time information of the subtitle text.

Abstract: A communication method includes receiving, from an indoor unit, an indoor unit model, an indoor unit protocol type, and a priority number; determining a communication protocol type and a communication address corresponding to communication with the indoor unit according to the indoor unit model, the indoor unit protocol type, and the priority number; and sending the communication protocol type and the communication address to the indoor unit for the indoor unit to establish a communicative connection to an outdoor unit according to the communication protocol type and the communication address.

Abstract: A method and apparatus for pushing subtitle data in a live scenario. The method includes: obtaining video stream data and audio stream data, the audio stream data being data corresponding to an audio part in the video stream data; generating the subtitle data according to the audio stream data, the subtitle data comprising a subtitle text corresponding to a speech in the audio stream data and time information of the subtitle text; and pushing, in response to pushing the video stream data to a user terminal, the subtitle data to the user terminal, the subtitle data instructing the user terminal to synchronously display the subtitle text with live pictures in the video stream data and the audio part in the audio stream data according to the time information of the subtitle text.

Abstract: Disclosed are a neural network training method, a neural network training device and an electronic device. The neural network training method includes: training a first neural network to be trained by using sample data; determining an indicator parameter of the first neural network in a current training process; determining an update manner corresponding to a preset condition if the indicator parameter meets the preset condition; and updating a parameter of a batch normalization layer in the first neural network based on the update manner. In this way, sparsing of a feature map output by a neural network is implemented, thereby reducing an amount of data to be transmitted and improving computation speed of a chip.

Abstract: Embodiments of the present systems and methods may provide techniques for systematic drug repositioning that fully leverages drug-drug interactions. In embodiments, the present systems and methods may generate a drug similarity metric based on DDI data and make drug repositioning predictions by leveraging the new similarity metric. For example, in an embodiment, a computer-implemented method for conducting a drug repositioning trial may comprise receiving data relating to drug-drug interactions, generating drug interaction vectors for each drug in the received data, generating a logistic regression model to generate drug repositioning hypotheses, generating at least one drug repositioning hypothesis using the logistic regression model, and conducting a drug repositioning trial using the generated drug repositioning hypothesis.

Abstract: Link prediction for biomedical entities. A neural network is trained using known associations between biomedical entities, including their vector representations and additional information-carrying content describing the biomedical entities. The trained network infers or predicts unobserved associations between two entities.

Abstract: Techniques regarding efficient means for chemical compound discovery are provided. For example, one or more embodiments can regard a system, which can comprise a memory that stores computer executable components and a processor, operably coupled to the memory, that can execute the computer executable components stored in the memory. The computer executable components can comprise a test component that can determine a first parameter value of a tested chemical compound from a plurality of chemical compounds. Additionally, a model component can generate a regression analysis model using a value information analysis. The regression analysis model can regard the plurality of chemical compounds based on the first parameter value. Further, an identification component can identify a preferred chemical compound from the plurality of chemical compounds based on the regression analysis model. A second parameter value of the preferred chemical compound can be greater than a defined threshold.

Abstract: Embodiments of the present invention disclose a method and an apparatus of establishing an image search relevance prediction model, and an image search method and apparatus. The method of establishing an image search relevance prediction model comprises: training a pre-constructed original deep neural network by using a training sample, wherein the training sample comprises: a query and image data, and the original deep neural network comprises: a representation vector generation network and a relevance calculation network; and using the trained original deep neural network as the image search relevance prediction model. The technical solution of the present invention optimizes the existing image search technology, and has stronger capabilities than the prior art as well as various integrations and variations in terms of semantic matching between a query and an image text, semantic matching between a query and image content, click generalization and the like.

Abstract: Disclosed is a method for convolution calculation in a neural network, comprising: reading an input feature map, depthwise convolution kernels and pointwise convolution kernels from a dynamitic random access memory (DRAM); performing depthwise convolution calculations and pointwise convolution calculations according to the input feature map, the depthwise convolution kernels and the pointwise convolution kernels to obtain output feature values of a first predetermined number p of points on all pointwise convolution output channels; storing the output feature values of a first predetermined number p of points on all pointwise convolution output channels into an on-chip memory, wherein the first predetermined number p is determined according to at least one of available space in the on-chip memory, a number of the depthwise convolution calculation units, and width, height and channel dimensions of the input feature map; and repeating the above operation obtain output feature values of all points on all pointwis