Abstract: A display panel, comprising a first insulating structural layer, a first crack detection line, a second insulating structural layer and a second crack detection line which are sequentially arranged on a substrate, wherein the first crack detection line and the second crack detection line are both located in a peripheral area and are arranged around a display area, one end of the first crack detection line is configured to receive a detection signal, and the other end of the first crack detection line is configured to output a first output signal, and one end of the second crack detection line is configured to receive a detection signal and the other end of the second crack detection line is configured to output a second output signal.

Abstract: Described herein are systems, methods, and other techniques for extracting one or more keyphrases from an input text. The input text may include a plurality of words. A plurality of token-level attention matrices may be generated using a transformer-based machine learning model. The plurality of token-level attention matrices may be converted into a plurality of word-level attention matrices. A set of candidate phrases may be identified from the plurality of words based on the plurality of word-level attention matrices. The one or more keyphrases may be selected from the set of candidate phrases.

Abstract: The present disclosure provides a color conversion substrate and manufacturing method thereof and a display panel. The color conversion substrate includes a base substrate; a color conversion layer on the base substrate and including a bank portion and a plurality of sub-portions having different colors, the bank portion is between adjacent sub-portions to separate the adjacent sub-portions and configured to absorb incident light, and the plurality of sub-portions having different colors are configured to convert the incident light in a same color into light having different colors; an anti-color-interference pattern on a side of the color conversion layer distal to the base substrate, an orthographic projection of the anti-color-interference pattern on the base substrate is within an orthographic projection of the bank portion on the base substrate, the anti-color-interference pattern is configured such that the incident light is refracted and then is transmitted into the bank portion.

Abstract: A reflective display panel and a display device are provided. The reflective display includes: a first liquid crystal cell including a first substrate and a second substrate oppositely arranged to each other, and a first liquid crystal layer between the first substrate and the second substrate; a plurality of pixel units on the first substrate, where each pixel unit includes a first sub-pixel unit and a second sub-pixel unit; an optical structure, arranged at a light-emitting side of the first liquid crystal cell and covering the pixel units.

Abstract: An photolithographic apparatus includes a particle removing cassette selectively extendable from the processing apparatus. The particle removing cassette includes a wind blade slit and an exhausting slit. The wind blade slit is configured to direct pressurized cleaning material to a surface of the mask to remove the debris particles from the surface of the mask. The exhausting slit collects the debris particles separated from the surface of the mask and contaminants through the exhaust line. In some embodiments, the wind blade slit includes an array of wind blade nozzles spaced apart within the wind blade slit. In some embodiments, the exhausting slit includes array of exhaust lines spaced apart within the exhausting slit.

Abstract: A display panel, a display module, and a display device are provided. The display panel includes a total reflection display unit and a frame enclosing the reflective display unit. The frame includes a light-shielding portion and a reflective layer. The light-shielding portion has a first surface parallel to a first direction and a second surface oppositely located to the first surface. The reflective layer covers the first surface or the second surface. The reflective layer can reflect the light irradiated on the surface of the light shielding portion or entering the light shielding portion back to the total reflection display unit again, avoiding the waste caused by the scattered light caused by the propagation path of the light, thereby improving the utilization rate of light.

Abstract: Methods and systems for training and/or using a language selection model for use in determining a particular language of a spoken utterance captured in audio data. Features of the audio data can be processed using the trained language selection model to generate a predicted probability for each of N different languages, and a particular language selected based on the generated probabilities. Speech recognition results for the particular language can be utilized responsive to selecting the particular language of the spoken utterance. Many implementations are directed to training the language selection model utilizing tuple losses in lieu of traditional cross-entropy losses. Training the language selection model utilizing the tuple losses can result in more efficient training and/or can result in a more accurate and/or robust model—thereby mitigating erroneous language selections for spoken utterances.

Abstract: A sunken wheel-side motor drive axle that includes an axle beam assembly, two drive motor assemblies, two reducer assemblies and two hub assemblies. The two drive motor assemblies, two reducer assemblies and two hub assemblies are respectively disposed at left and right ends of the axle beam assembly. Moreover, the drive motor assemblies are symmetrically disposed at a bottom of the left and right ends of the axle beam assembly respectively; and the drive motor assemblies transmit power to the reducer assemblies, and the reducer assemblies drive the hub assemblies to rotate.

Abstract: Disclosed is a method for training an expression transfer model performed by a computer device, the method including: obtaining a source domain facial image of a first object, a target domain facial image of a second object and a facial feature image associated with the target domain facial image; applying the facial feature image and the source domain facial image to an expression transfer model to obtain a synthesized facial image of the first object; applying the synthesized facial image and the target domain facial image to a discriminative network model obtain two discrimination results; applying the synthesized facial image and the target domain facial image to an image classification model to obtain a category feature vector for identifying a difference between the synthesized facial image and the target domain facial image; and updating the expression transfer model according to the category feature vector and the discrimination results.

Abstract: Disclosed in embodiments of the present disclosure are methods and apparatuses for realizing a globally unique index, which are applicable to a distributed database. One of the methods includes: receiving a structured query language, SQL, request, obtaining a conflict detection SQL statement corresponding to the SQL request and calculating an expected value; issuing the conflict detection SQL statement to each database node for execution; obtaining an execution result of the conflict detection SQL statement, and comparing the execution result of the conflict detection SQL statement with the expected value; and when the execution result of the conflict detection SQL statement is less than or equal to the expected value, processing an original SQL statement of the SQL request, and issuing the original SQL statement to the each database node for execution.

Abstract: Development of resists are useful, for example, to form a patterning mask in the context of high-resolution patterning. Development can be accomplished using a halide-containing chemistry such as a hydrogen halide. A metal-containing resist film may be deposited on a semiconductor substrate using a dry or wet deposition technique. The resist film may be an EUV-sensitive organo-metal oxide or organo-metal-containing thin film resist. After exposure, the photopatterned metal-containing resist is developed using wet or dry development.

Abstract: The invention provides a display device including a display panel and a backlight module including a light guide plate, a light source, and an optical film. The light guide plate has light incident and exit surfaces. The light source is disposed at one side of the light incident surface. The optical film is overlapped with the light exit surface, and has first optical microstructures facing the light exit surface. The display panel includes a liquid crystal cell overlapped with the light exit surface, first and second polarizers respectively disposed at two opposite sides of the liquid crystal cell, and a phase retardation film disposed between the first and second polarizers. The first polarizer is located between the liquid crystal cell and the optical film. An axial direction of an optical axis of the phase retardation film is perpendicular to an axial direction of an absorption axis of the first polarizer.

Abstract: A tiling display system has multiple display panels. The display panels may be positioned by positioners that are coupled to the display panels. In an untiled operating mode, the display panels of the tiling display system are moved apart. In this mode, each display panel can display different content such as different static images of artwork. In a tiled operating mode, the display panels of the tiling system are moved together to form a single display that displays a single still or moving image. Components on the backs of the display panels and/or along the edges of the display panels may be used to facilitate panel-to-panel alignment, to help couple adjacent panels together, and to support the transfer of power and/or data signals among the panels. The components may include optical and/or electrical alignment sensors, magnets for alignment and coupling, and transmitters and receivers for transmitting and receiving signals.

Abstract: A semiconductor structure and a manufacturing method thereof are provided. The semiconductor structure includes a first under-bump metallization (UBM) pattern, a first conductive via, and a first dielectric layer laterally covering the first UBM pattern and the first conductive via. Entireties of a top surface and a bottom surface of the first UBM pattern are substantially planar. The first conductive via landing on the top surface of the first UBM pattern includes a vertical sidewall and a top surface connected to the vertical sidewall, and a planarized mark is on the top surface of the first conductive via. A bottom surface of the first dielectric layer is substantially flush with the bottom surface of the first UBM, and a top surface of the first dielectric layer is substantially flush with the top surface of the first conductive via.

Abstract: A repairing method for an array substrate is provided. The array substrate includes at least one defective signal line, which is a data line or a gate line having a breakpoint. The defective signal line is divided by the breakpoint into a first portion and a second portion. The repairing method includes disconnecting a connection between a first thin film transistor and a data line or a gate line to which the first thin film transistor is connected, the first thin film transistor being a thin film transistor closest to the breakpoint among thin film transistors connected to the first portion; electrically connecting a first terminal of the first thin film transistor to the first portion; and taking a common electrode line segment from a common electrode line to which the first pixel electrode corresponds as a repair line.

Abstract: A spectrally-resolved Raman water lidar, including: a transmitter unit, a receiver unit, and a data acquisition and control unit. The transmitter unit includes a seeder, a solid Neodymium-doped Yttrium Aluminum Garnet (Nd:YAG) laser, a beam expander, and a first reflecting mirror to emit a 354.8-nm laser beam. The receiver unit includes a telescope, an iris, a collimator, a second reflecting mirror, a first bandpass filter, a beam splitter, a narrow-band interference filter, a third lens, a first detector, a second bandpass filter, a coupler and a home-made dual-grating polychromator to enable simultaneous profiling of backscattered Raman spectrum signals from water vapor, water droplets and ice crystals as well as aerosol fluorescence in the atmosphere. The data acquisition and control unit includes a computer to store the acquired data and guarantee an automatic operation of the lidar system through a time-sequence circuit.

Abstract: A backlight module including a light guide plate, a light source, an optical film, a diffusion sheet, a first prism sheet and a second prism sheet is provided. The light guide plate has a light incident surface and a light emitting surface connected to each other. The light source is disposed on a side of the light incident surface. The optical film is overlapped with the light emitting surface and has a plurality of optical microstructures facing the light emitting surface. The diffusion sheet is disposed between the light guide plate and the optical film. The first prism sheet and the second prism sheet are overlapped with the optical film and are positioned on a side of the optical film being far away from the light guide plate. A display apparatus adopting the backlight module is also provided.

Abstract: A method of forming a high electron mobility transistor (HEMT) includes a first III-V compound layer and a second III-V compound layer disposed on the first III-V compound layer and is different from the first III-V compound layer in composition. A source feature and a drain feature are disposed on the second III-V compound layer. A p-type layer is disposed on a portion of the second III-V compound layer between the source feature and the drain feature. A gate electrode is disposed on the p-type layer. A capping layer is disposed on the second III-V compound layer.

Abstract: An embodiment of this application discloses an artificial intelligence-based image generation method performed by a computer device. The method includes: acquiring a source image including a target object whose pose is to be transformed, and a target image including a reference object presenting a target pose; determining a pose transition matrix according to a model pose corresponding to the pose of the target object and a model pose corresponding to the target pose of the reference object; extracting a basic appearance feature of the target object from the source image; processing the basic appearance feature based on the pose transition matrix, to obtain a target appearance feature of the target object in the target pose; and generating a target synthetic image of the target object in the target pose based on the target appearance feature.

Abstract: Embodiments disclosed herein provide for systems and methods for implementing chaos engineering in distributed systems with chaos engineering trials. The systems and methods provide for a RESTful API that is configured to generate a chaos engineering experiment including a plurality of steps that test one or more conditions in the distributed system as well as generate a chaos engineering trial based on the chaos engineering experiment and one or more parameters.