Abstract: An operating circuit and an operating method of a resistive random-access memory are provided, the operating circuit includes: at least one capacitance connected in series with the resistive random-access memory, so that the resistive random-access memory is grounded through the at least one capacitance. The operating method includes: connecting at least one capacitance in series with a resistive random-access memory, so that the resistive random-access memory is grounded through the capacitance; applying a forming pulse voltage or a set pulse voltage on the resistive random-access memory to achieve a forming operation or a set operation of the resistive random-access memory.

Abstract: The present disclosure provides a complementary phototransistor pixel unit, a sensing and computing array structure and an operation method thereof. The complementary phototransistor pixel unit includes: a first photoelectric field effect transistor, which is a photoelectric field effect transistor based on an ultra-thin body and buried oxide layer; and a second photoelectric field effect transistor, the second photoelectric field effect transistor is a photoelectric field effect transistor based on an ultra-thin body and buried oxide layer, each of the first photoelectric field effect transistor and the second photoelectric field effect transistor is four-end device and has a gate electrode G, a source electrode S, a drain electrode D, and a well base electrode B, and the source electrode S or drain electrode D of the first photoelectric field effect transistor is connected to the source electrode S or drain electrode D of the second photoelectric field effect transistor.

Abstract: A platinum catalyst recycling device with multiple filtration structures is provided. The device includes a support base, a recycling housing, and a crushing housing. A top of the support base is fixedly connected to the recycling housing, a top surface of the recycling housing is fixedly connected to the crushing housing, and a feed port is provided on a top surface of the crushing housing. A feed tube is installed at the feed port of the crushing housing, and a crushing roller assembly is installed in an inner cavity of the crushing housing. The crushing roller assembly is disposed in an inner cavity of a crushing chamber, and the crushing chamber is fixedly connected to a sidewall of the inner cavity of the crushing housing. A bottom of the crushing chamber is fixedly connected to a guide plate. Uniform plates are disposed below the guide plate.

Abstract: The present disclosure provides a method and an apparatus for operating an in-memory computing architecture applied to a neural network and a device, the method includes: generating a mono-pulse input signal based on discrete time coding; inputting the mono-pulse input signal into a memory array of the in-memory computing architecture to generate a bit line current signal corresponding to the memory array; and controlling a neuron circuit of the in-memory computing architecture to output a mono-pulse output signal based on discrete time coding according to the bit line current signal, wherein the mono-pulse output signal is configured as a mono-pulse input signal of a memory array of the next layer of neural network in the next in-memory computing cycle.

Abstract: The present application discloses a UTBB photodetector pixel unit, array and method, including: a silicon film layer, a buried oxide layer, a charge collection layer and a substrate, the silicon film layer, the buried oxide layer, the charge collection layer and the substrate being arranged in sequence from top to bottom; the silicon film layer includes NMOS transistors or PMOS transistors; the charge collection layer includes charge collection control regions and charge accumulation regions; and the substrate includes an N-type substrate or a P-type substrate. A centripetal electric field is formed around the charge accumulation regions, and photo-generated charges are accumulated in the corresponding pixel units under the action of the centripetal electric field.

Abstract: An eaves lamp structure with a segmented protective cover includes a base, a lamp string, and a shielding cover. Two sides of the base are extended upwards with flexible clamping arms, and the two flexible clamping arms enclose a clamping cavity; the lamp string comprises multiple lamp bodies and a cable connecting adjacent two lamp bodies, the shape of the lamp body is suitable for the clamping cavity and can be detachably installed between the two flexible clamping arms; the shielding cover can be detachably installed in the clamping cavity and arranged between the two lamp bodies, the cable is shielded inside the shielding cover, and two sides of the shielding cover are in contact with the side ends of the lamp body and limit the lamp body. By setting the flexible clamping arm on the base, the installation of the lamp string and shielding cover is achieved.

Abstract: The present disclosure provides an image sensing computing unit and its operating method, an image sensing computer and an electronic device. Among them, the image sensing computing unit includes a first photosensitive unit and a second photosensitive unit. The second photosensitive unit is connected in series with the first photosensitive unit. The changing direction of the first threshold voltage of the first photosensitive unit when receiving light is opposite to the changing direction of the second threshold voltage of the second photosensitive unit when receiving light, so as to implement an in-situ logical operation between light input signals.

Abstract: A method for selection of a calibration set and a validation set based on spectral similarity and modeling. The method includes: performing NIR spectrometry on original samples to obtain a spectral matrix of the original samples; randomly selecting m samples as an independent test set; calculating spectral similarity between each of the samples in the independent test set and each of the remaining samples in the original samples respectively to obtain g samples having the highest similarity to be written into the validation set; and calculating spectral similarity between each of the samples in the validation set and each of the remaining samples in the original samples respectively to obtain n samples having the highest similarity to be written into the calibration set. Based on the validation set and the calibration set selected through the method, an obtained model can predict unknown samples more accurately.

Abstract: A system and a method for compressing an image based on a FLASH in-memory computing array are provided. The system includes: a convolutional neural network for encoding of the FLASH in-memory computing array, a convolutional neural network for decoding based on the FLASH in-memory computing array, and a quantization module; the convolutional neural network for encoding based on the FLASH in-memory computing array is configured to encode an original image to obtain a feature image; the quantization module is configured to quantize the feature image to obtain a quantized image; the convolutional neural network for decoding based on the FLASH in-memory computing array is configured to decode the quantized image to obtain a compressed image.

Abstract: A method of solving a partial differential equation based on a non-volatile memory array includes converting a to-be-solved partial differential equation into an iterative relation, selecting a reusable sub-matrix cell from the iterative coefficient matrix, and storing the sub-matrix cell in the memory array, extracting an input vector from the iteration vector, inputting the input vector into the memory array, updating a portion of the iteration vector by adding an obtained output vector to a portion of the constant vector, extracting the input vector from an updated iteration vector again, and inputting the input vector into the memory array until all elements of the iteration vector are updated to obtain an iteration vector for a next iteration, and ending the iteration when a preset number of iterations is reached or an error is less than a preset range.

Abstract: The technology described herein provides an automated software-testing platform that uses reinforcement learning to discover how to perform tasks used in testing. The technology described herein is able to perform quality testing even when prescribed paths to completing tasks are not provided. The reinforcement-learning agent is not directly supervised to take actions in any given situation, but rather learns which sequences of actions generate the most rewards through the observed states and rewards from the environment. In the software-testing environment, the state can be user interface features and actions are interactions with user interface elements. The testing system may recognize when a sought after state is achieved by comparing a new state to a reward criteria.

Abstract: The present application discloses a pixel unit and a signal processing method for a pixel unit. The pixel unit includes at least one pixel, and the pixel includes: an N-type main pixel, a P-type main pixel, and a sub-pixel; and the sub-pixel is located between the N-type main pixel and the P-type main pixel; or the pixel includes at least a first pixel and a second pixel that are adjacent to each other; the first pixel includes an N-type main pixel, and the second pixel includes a P-type main pixel; the first pixel and the second pixel share one sub-pixel; the sub-pixel is configured to generate and output a signal difference between the N-type main pixel and the P-type main pixel according to the current.

Abstract: The technology described herein provides a cloud reinforcement-learning architecture that allows a single reinforcement-learning model to interact with multiple live software environments. The live software environments and the single reinforcement-learning model run in a distributed computing environment (e.g., cloud environment). The single reinforcement-learning model may run on a first computing device(s) with a graphical processing unit (GPU) to aid in training the single reinforcement-learning model. At a high level, the single reinforcement-learning model may receive state telemetry data from the multiple live environments. The single reinforcement-learning model selects an available action for each set of state telemetry data received and communicates the selection to appropriate the test agent. The test agent then facilitates completion of the action within the software instance being tested in the live environment. A reward is then determined for the action.

Abstract: A BVP10 protein shown in SEQ ID NO:2 for controlling tetranychid mites and use of the protein is provided. The BVP10 protein has a median lethal concentration of 19.07 ?g/mL against Tetranychus urticae, a median lethal concentration of 58.05 ?g/mL against Panonychus citri, a median lethal concentration of 36.08 ?g/mL against Tetranychus cinnabarinus, and a control effect of 79.53%-95.45% against strawberry red spider mites. The protein is provided for preparing a novel mite pesticide.

Abstract: A spray device includes a spray assembly. The spray assembly is configured to spray cleaning liquid towards a to-be-cleaned surface of a to-be-processed workpiece. The spray assembly includes a spray head and multiple liquid inlet pipelines. The spray assembly is configured to spray cleaning liquids with different concentrations, temperatures, and/or flow rates corresponding to different positions in a radial direction of the to-be-cleaned surface to cause a cleaning rate of the cleaning liquids at the different positions in the radial direction of the to-be-cleaned surface to be consistent.

Abstract: The technology described herein trains a reinforcement-learning model in a simulated environment. A simulated environment contrasts with a live environment. A live environment is a computing environment with which the reinforcement-learning model will interact once it is deployed. In order to be effective, the simulated environment may provide inputs to the reinforcement-learning model in the same format as the reinforcement-learning model receives from the live environment. In aspects, the training in the simulated environment may act as pre-training for training in the live environment. Once pre-trained, the reinforcement-learning model may be deployed in a live environment and continue to learn how to perform the same task in different ways, learn how to perform additional tasks, and/or improve performance of a task learned in pre-training. In aspects, the reinforcement-learning model may be used to discover unhealthy conditions in software by performing the tasks it has learned.

Abstract: The technology described herein provides an automated software-testing platform that uses reinforcement learning to discover how to perform tasks used in testing. The technology described herein is able to perform quality testing even when prescribed paths to completing tasks are not provided. The reinforcement-learning agent is not directly supervised to take actions in any given situation, but rather learns which sequences of actions generate the most rewards through the observed states and rewards from the environment. In the software-testing environment, the state can be user interface features and actions are interactions with user interface elements. The testing system may recognize when a sought after state is achieved by comparing a new state to a reward criteria.

Abstract: The technology described herein provides a cloud reinforcement-learning architecture that allows a single reinforcement-learning model to interact with multiple live software environments. The live software environments and the single reinforcement-learning model run in a distributed computing environment (e.g., cloud environment). The single reinforcement-learning model may run on a first computing device(s) with a graphical processing unit (GPU) to aid in training the single reinforcement-learning model. At a high level, the single reinforcement-learning model may receive state telemetry data from the multiple live environments. The single reinforcement-learning model selects an available action for each set of state telemetry data received and communicates the selection to appropriate the test agent. The test agent then facilitates completion of the action within the software instance being tested in the live environment. A reward is then determined for the action.

Abstract: A BVP8 protein for killing tetranychid mites and use thereof are provided. The protein is as set forth in SEQ ID NO. 2. The BVP8 protein has a median lethal concentration of 12.98 ?g/mL against Tetranychus urticae, 33.45 ?g/mL against Panonychus citri, and 26.32 ?g/mL against Tetranychus cinnabarinus, and shows an inhibitory effect against the hatching and cleavage of Tetranychus urticae eggs, with the egg cleavage rate of 75.86% after 72 h. The protein provides a new option for the preparation of a novel miticide.

Abstract: A polymer composition containing an impact modifier and a fibrous filler dispersed within a polymer matrix is provided. The polymer matrix contains a high performance thermoplastic polymer that exhibits a deflection temperature under load of about 40° C. or more as determined in accordance with ISO 75-2:2013 at a load of 1.8 MPa and a melting temperature of about 140° C. or more. The polymer composition exhibits a thermal shock resistance value of about 800 or more.