Abstract: A system and a method to train a neural network are disclosed. A first image is weakly and strongly augmented. The first image, the weakly and strongly augmented first images are input into a feature extractor to obtain augmented features. Each weakly augmented first image is input to a corresponding first expert head to determine a supervised loss for each weakly augmented first image. Each strongly augmented first image is input to a corresponding second expert head to determine a diversity loss for each strongly augmented first image. The feature extractor is trained to minimize the supervised loss on weakly augmented first images and to minimize a multi-expert consensus loss on strongly augmented first images. Each first expert head is trained to minimize the supervised loss for each weakly augmented first image, and each second expert head is trained to minimize the diversity loss for each strongly augmented first image.

Abstract: Provided is an input voltage endurance protection architecture applied to a high-voltage operational amplifier with high input amplitude and high linearity. The input voltage endurance protection architecture includes three parts: a main operational amplifier, an auxiliary operational amplifier and an input stage voltage endurance protection circuit. The main operational amplifier is a high-voltage general-purpose operational amplifier, the auxiliary operational amplifier is a single-stage differential amplifier, and the single-stage differential operational amplifier is connected to a degeneration resistor Rbias. In addition, the auxiliary operational amplifier has a same connection method as the main operational amplifier at a positive input terminal and a negative input terminal, and both the positive input terminal and the negative input terminal are protected by an input stage voltage endurance protection circuit and receive and process input signals simultaneously.

Abstract: Provided is an input voltage endurance protection architecture applied to a high-voltage operational amplifier with high input amplitude and high linearity. The input voltage endurance protection architecture includes three parts: a main operational amplifier, an auxiliary operational amplifier and an input stage voltage endurance protection circuit. The main operational amplifier is a high-voltage general-purpose operational amplifier, the auxiliary operational amplifier is a single-stage differential amplifier, and the single-stage differential operational amplifier is connected to a degeneration resistor Rbias. In addition, the auxiliary operational amplifier has a same connection method as the main operational amplifier at a positive input terminal and a negative input terminal, and both the positive input terminal and the negative input terminal are protected by an input stage voltage endurance protection circuit and receive and process input signals simultaneously.

Abstract: A method and system for training a neural network are provided. The method includes receiving an input image, selecting at least one data augmentation method from a pool of data augmentation methods, generating an augmented image by applying the selected at least one data augmentation method to the input image, and generating a mixed image from the input image and the augmented image.

Abstract: A method of making a plasmonic metal/graphene heterostructure comprises heating an organometallic complex precursor comprising a metal at a first temperature T1 for a first period of time t1 to deposit a layer of the metal on a surface of a heated substrate, the heated substrate in fluid communication with the precursor; and heating, in situ, the precursor at a second temperature T2 for a second period of time t2 to simultaneously form on the layer of the metal, a monolayer of graphene and a plurality of carbon-encapsulated metal nanostructures comprising the metal, thereby providing the plasmonic metal/graphene heterostructure. The heated substrate is characterized by a third temperature T3. The plasmonic metal/graphene heterostructures, devices incorporating the heterostructures, and methods of using the heterostructures are also provided.

Abstract: A method for computing a dominant class of a scene includes: receiving an input image of a scene; generating a segmentation map of the input image, the segmentation map being labeled with a plurality of corresponding classes of a plurality of classes; computing a plurality of area ratios based on the segmentation map, each of the area ratios corresponding to a different class of the plurality of classes of the segmentation map; and outputting a detected dominant class of the scene based on a plurality of ranked labels based on the area ratios.

Abstract: Some aspects of embodiments of the present disclosure relate to using a boundary aware loss function to train a machine learning model for computing semantic segmentation maps from input images. Some aspects of embodiments of the present disclosure relate to deep convolutional neural networks (DCNNs) for computing semantic segmentation maps from input images, where the DCNNs include a box filtering layer configured to box filter input feature maps computed from the input images before supplying box filtered feature maps to an atrous spatial pyramidal pooling (ASPP) layer. Some aspects of embodiments of the present disclosure relate to a selective ASPP layer configured to weight the outputs of an ASPP layer in accordance with attention feature maps.

Abstract: A method for recognizing a character includes: obtaining a character; converting the character into a radical based character recognition result, where the radical based character recognition result comprises symbols indicating radicals of the character and a structure of the radicals of the character; and recognizing the character based on the radical based character recognition result.

Abstract: A method for computing a dominant class of a scene includes: receiving an input image of a scene; generating a segmentation map of the input image, the segmentation map including a plurality of pixels, each of the pixels being labeled with a corresponding class of a plurality of classes; computing a plurality of area ratios based on the segmentation map, each of the area ratios corresponding to a different class of the plurality of classes of the segmentation map; applying inference to generate a plurality of ranked labels based on the area ratios; and outputting a detected dominant class of the scene based on the plurality of ranked labels.

Abstract: A switch mode power supply includes a flyback power converter and a control circuit. The flyback power converter includes an input, an output, a transformer coupled between the input and the output, a power switch coupled between the input and the transformer, and a clamp circuit coupled between the input and the transformer. The clamp circuit includes a capacitor and a clamp switch coupled in series with the capacitor. The control circuit is configured to control the power switch and the clamp switch. The switch mode power supply further includes at least one additional capacitor coupled in parallel with the capacitor of the clamp circuit to facilitate selection of a combination of capacitors to adjust a resonant frequency of the clamp switch for optimizing efficiency of the power supply. Other examples switch mode power supplies and/or methods for adjusting a resonant frequency of flyback power converters are also disclosed.

Abstract: A process for improving the performance of the sliding rheostat of 5G communication high-frequency signal board with the sliding rheostat slides along between two bonding pads, includes the following steps: outer layer etching; resin plugging: a. plugging the resinous ink into the pre-plugging position; b: baking, baking on the baking plate of the oven after the plugging is finished; board polishing: using a ceramic brush to process the plugged board, then using a non-woven fabric blush to polish the surface that is polished by ceramic brush. The present invention provides a process for improving the performance of the sliding rheostat of 5G communication high-frequency signal board. The resin plugging method is used to plug the gap between the conductors of the sliding rheostat, so as to prevent the sliding rheostat from being unable to slide due to the altitude difference between conductors of the high-frequency signal board.

Abstract: A method for computing a dominant class of a scene includes: receiving an input image of a scene; generating a segmentation map of the input image, the segmentation map including a plurality of pixels, each of the pixels being labeled with a corresponding class of a plurality of classes; computing a plurality of area ratios based on the segmentation map, each of the area ratios corresponding to a different class of the plurality of classes of the segmentation map; applying inference to generate a plurality of ranked labels based on the area ratios; and outputting a detected dominant class of the scene based on the plurality of ranked labels.

Abstract: A method for recognizing a character includes: obtaining a character; converting the character into a radical based character recognition result, where the radical based character recognition result comprises symbols indicating radicals of the character and a structure of the radicals of the character; and recognizing the character based on the radical based character recognition result.

Abstract: According to some aspects of the present disclosure, switched-mode power converters and control methods are disclosed. Example power converters include an input terminal, an output terminal, and a transformer coupled between the input terminal and the output terminal. The transformer includes a primary winding and a secondary winding. The power supply also includes at least one switch coupled to the primary winding, and a primary controller coupled to the at least one switch to control switching operation of the at least one switch. The primary controller includes a voltage reference. The power supply further includes a secondary controller configured to detect a no load condition at the output terminal, and operate the primary controller in a burst mode by adjusting the voltage reference of the primary controller when the secondary controller detects the no load condition at the output terminal.

Abstract: A power factor correction (PFC) stage of a power supply unit has a PFC circuit including a rectifier circuit, a PFC controller circuit with a PFC switch, a current sensor connected to the PFC switch, a high frequency bypass capacitor connected between the PFC controller circuit and the rectifier circuit, and a bulk storage capacitor connected between the PFC controller circuit and an output of the PFC stage. The PFC stage also has a negative temperature coefficient thermistor connected in series with the PFC switch and the current sensor. During a start-up of the power supply unit, the PFC controller circuit causes the PFC switch to turn-on until the PFC controller circuit causes the PFC switch to turn-off after a current through the current sensor is sensed as being equal to or greater than a preset value.

Abstract: An intelligent scoring method and system for a text objective question, the method comprising: acquiring an answer image of a text objective question (101); segmenting the answer image to obtain one or more segmentation results of an answer string to be identified (102); determining whether any of the segmentation results has the same number of characters as the standard answer (103); if no, the answer is determined to be wrong (106); otherwise, calculating identification confidence of the segmentation result having the same number of words as the standard answer, and/or calculating the identification confidence of respective characters in the segmentation result having the same number of words as the standard answer (104); determining whether the answer is correct according to the calculated identification confidence (105). The method can automatically score text objective questions, thus reducing consumption of human resource, and improving scoring efficiency and accuracy.

Abstract: An optoelectronic device comprises a nanocomposite comprising a carbon nanostructure having a surface and a biomolecule adsorbed on the surface and forming a heterojunction at the interface of the carbon nanostructure and the biomolecule, the carbon nanostructure and the biomolecule each characterized by respective conduction band edges and valence band edges. The device further comprises first and second electrodes in electrical communication with the nanocomposite. The conduction band edge offset, the valence band edge offset, or both, across the heterojunction is greater in energy than the binding energy of an exciton generated in the carbon nanostructure or the biomolecule upon the absorption of light such that the exciton dissociates at the heterojunction to an electron, which is injected into one of the carbon nanostructure and the biomolecule, and a hole, which is injected into the other of the carbon nanostructure and the biomolecule.

Abstract: A gold-plating etching process for 5G high-frequency signal boards is carried out according to the following steps: the outer dry film, the plug gold-plating, the film removing, and the alkaline etching. The alkaline etching solution comprises 100 to 150 g/L of cupric chloride, 90 to 120 g/L of ammonium chloride, and ammonia. The pH value is 9.6 to 9.8. The ratio of the ammonia and the alkaline etching solution is (550-800):1000. The present invention provides a gold-plating etching process of 5G high-frequency signal boards. The alkaline etching procedure is performed right after gold-plating, eliminating the outer etching process after gold-plating. Costs of the outer film pressing, the exposure, and the development can be saved. The flow rate is improved. Requirements of 5G communication circuit boards are satisfied. That is, the transmission speed of 5G communication high-frequency signal boards of the present invention is fast.

Abstract: A switch mode power supply includes a flyback power converter and a control circuit. The flyback power converter includes an input, an output, a transformer coupled between the input and the output, a power switch coupled between the input and the transformer, and a clamp circuit coupled between the input and the transformer. The clamp circuit includes a capacitor and a clamp switch coupled in series with the capacitor. The control circuit is configured to control the power switch and the clamp switch. The switch mode power supply further includes at least one additional capacitor coupled in parallel with the capacitor of the clamp circuit to facilitate selection of a combination of capacitors to adjust a resonant frequency of the clamp switch for optimizing efficiency of the power supply. Other examples switch mode power supplies and/or methods for adjusting a resonant frequency of flyback power converters are also disclosed.

Abstract: According to some aspects of the present disclosure, switched-mode power converters and control methods are disclosed. Example power converters include an input terminal, an output terminal, and a transformer coupled between the input terminal and the output terminal. The transformer includes a primary winding and a secondary winding. The power supply also includes at least one switch coupled to the primary winding, and a primary controller coupled to the at least one switch to control switching operation of the at least one switch. The primary controller includes a voltage reference. The power supply further includes a secondary controller configured to detect a no load condition at the output terminal, and operate the primary controller in a burst mode by adjusting the voltage reference of the primary controller when the secondary controller detects the no load condition at the output terminal.