Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for realizing a multimodal image classifier. In an aspect, a method includes, for each image of a plurality of images: processing the image by a textual generator model to obtain a set of phrases that are descriptive of the content of the image, wherein each phrase is one or more terms, processing the set of phrases by a textual embedding model to obtain an embedding of predicted text for the image, and processing the image using an image embedding model to obtain an embedding of image pixels of the image. Then a multimodal image classifier is trained on the embeddings of predicted text for the images and the embeddings of image pixels for the images to produce, as output, labels of an output taxonomy to classify an image based on the image as input.

Abstract: An optical element driving mechanism is used for accommodating a first optical element and includes a fixed assembly, a movable part and a driving assembly. The movable part is configured to connect a second optical element. The second optical element corresponds to the first optical element. The movable part is movable relative to the fixed assembly. The driving assembly is configured to drive the movable part to move relative to the fixed assembly. The fixed assembly includes a first accommodating space configured to accommodate the first optical element.

Abstract: Embodiment of the present application discloses an array substrate and a display panel. An active layer includes a first connection segment, a second connection segment, and a third connection segment connected in sequence. A part of the second connection segment of the active layer overlaps with a data line, and another part of the second connection segment locates between two adjacent data lines. The third connection segment locates between the two adjacent data lines. A drain electrode contacts a part of the active layer exposed between the two adjacent data lines.

Abstract: A charging device interface circuit and a charging device, and the charging device interface circuit includes an AC input interface, a differential mode protection unit, a common mode protection unit and an AC/DC power conversion unit; the AC input interface is configured to be connected to alternating current; the differential mode protection unit includes a first-stage differential mode protection unit and a second-stage differential mode protection unit; the first-stage differential mode protection unit is connected to the AC input interface; the second-stage differential mode protection unit is connected to the first-stage differential mode protection unit; the common mode protection unit is connected in parallel with the second-stage differential-mode protection unit; an AC end of the AC/DC power conversion unit is connected to the common mode protection unit; and a DC end of the AC/DC power conversion unit is connected to an apparatus to be charged.

Abstract: Aspects of the present disclosure provide an automated labeling system. For example, the automated labeling system can include an automated labeling module (ALM) configured to receive wireless signals and ground truth of learning object and label the wireless signals with the ground truth when receiving the ground truth to generate labeled training data. The automated labeling system can also include a training database coupled to the ALM. The training database can be configured to store the labeled training data.

Abstract: A semiconductor device includes a logic circuit region having at least one core device and at least one input/output (I/O) device. The at least one core device has a first accumulative antenna ratio, and the at least one I/O device has a second accumulative antenna ratio. The first accumulative antenna ratio is greater than the second accumulative antenna ratio.

Abstract: The present disclosure discloses a pixel circuit and a display panel. The pixel circuit includes a light emitting module and a driving module including at least two driving units connected in parallel. a quantity of displayable gray scales can be improved or increased by configuring at least two driving units connected in parallel in the driving module where the at least two driving units can each correspondingly configure a light emitting current so that a plurality of light emitting brightness of the light emitting module can be implemented by a sum of these light emitting currents.

Abstract: A data processing method, a data processing apparatus, an electronic device, and a computer-readable storage medium are provided. The data processing method includes: in response to an operation of importing a resource file of an object to be tried on, performing processing on the resource file and obtaining a target file in a file format adapted to an effect editor; adding a rendering component to a preset node of the effect editor in response to an operation of dragging the target file to the preset node; and rendering, based on the rendering component, a try-on image in which a model of the object to be tried on is tried on a model of a preset part.

Abstract: A semiconductor device including a housing, a semiconductor chip disposed within the housing and having first and second metal electrodes, a first lead frame having a first end extending out of the housing and a second end terminating in a die pad, a top surface of the die pad including a cavity having a first quantity of solder disposed therein for electrically connecting the die pad to the first metal electrode, a second lead frame having a first end extending out of the housing and having a second end disposed adjacent the semiconductor chip, and a clip having a first end connected to the second of the lead frame and a second end extending over the semiconductor chip, a bottom surface of the second end of the clip including a recess having a second quantity of solder disposed therein for electrically connecting the clip to the second metal electrode.

Abstract: A method for training a machine learning model is described, comprising receiving, for each perturbation of a plurality of perturbations of model parameters of a starting version of the machine learning model, a change of loss of the machine learning model caused by the perturbation for a set of training data determined by feeding the set of training data to one or more perturbed versions of the machine learning model, estimating a gradient of the loss of the machine learning model with respect to the model parameters from the determined changes of loss and updating the starting version of the machine learning model to an updated version of the machine learning model by changing the model parameters in a direction for which the estimated gradient indicates a reduction of loss.

Abstract: An acoustic device includes a first sound producing component and a back cavity structure. The first sound producing component has a first front side and a first back side, wherein the first sound producing component is a high frequency sound unit, and the first front side faces a sound propagating opening of the acoustic device. The back cavity structure is connected to the first back side of the first sound producing component. The first sound producing component produces a first acoustic wave from the first front side towards the sound propagating opening, and the first sound producing component produces a second acoustic wave from the first back side towards a back cavity of the back cavity structure. The back cavity structure is configured to flatten a peak or a dip of a frequency response of the first sound producing component.

Abstract: An antenna module includes a first metal plate and a frame body. The frame body surrounds the first metal plate. The frame body includes a first antenna radiator, a second antenna radiator, a third antenna radiator, a first breakpoint and a second breakpoint. The first antenna radiator includes a first feeding end and excites a first frequency band. The second antenna radiator includes a second feeding end and excites a second frequency band. The third antenna radiator includes a third feeding end and excites a third frequency band. The first breakpoint is located between the first antenna radiator and the second antenna radiator. The second breakpoint is located between the second antenna radiator and the third antenna radiator. An electronic device including the above-mentioned antenna module is also provided.

Abstract: A transformer includes a bobbin and a plurality of coils wound on the bobbin. The plurality of coils includes a first primary coil; a second primary coil, located above the first primary coil and electrically connected to the first primary coil; a secondary coil, located between the first primary coil and the second primary; a first auxiliary coil, located above the second primary coil; and a second auxiliary coil, located on the first auxiliary coil and electrically connected to the first auxiliary coil.

Abstract: Composite dielectric structures for semiconductor die assemblies, and associated systems and methods are disclosed. In some embodiments, the composite dielectric structure includes a flexible dielectric layer configured to conform to irregularities (e.g., particles, defects) at a bonding interface of directly bonded semiconductor dies (or wafers). The flexible dielectric layer may include a polymer material configured to deform in response to localized pressure generated by the irregularities during bonding process steps. The composite dielectric structure includes additional dielectric layers sandwiching the flexible dielectric layer such that the composite dielectric structure can provide robust bonding strength to other dielectric layers through the additional dielectric layers. In some embodiments, a chemical vapor deposition process may be used to form the composite dielectric structure utilizing siloxane derivatives as a precursor.

Abstract: Provided a separator, a preparation method thereof, and a secondary battery, a battery module, a battery pack, and an electric apparatus containing such separator. The separator includes: a substrate and a coating applied on at least one surface of the substrate, where the coating includes 40 wt % to 90 wt % first organic particles, 2 wt % to 15 wt % pressure-sensitive binder, and optionally 0 wt % to 50 wt % second organic particles. The first organic particles include an organic polymer and an inorganic substance. The pressure-sensitive binder includes an organic polymer and a plasticizer. The separator has good ionic conductivity and adhesion performance at room temperature, is not adhered under a pressure of 1 MPa or below and is significantly adhered under a pressure of 2 MPa or above, which can significantly improve structural stability and ionic conductivity of an electrochemical device while satisfying requirements for yield rate of production and kinetic performance of a battery cell.

Abstract: A semiconductor device structure includes a semiconductor substrate, an active region, a STI (shallow trench isolation) region, and an interconnection layer. The semiconductor substrate has an original surface. The active region is within the semiconductor substrate, wherein the active region includes a transistor and the transistor includes a gate structure with a bottom surface under the original surface, a first conductive region, and a second conductive region. The STI region surrounds the active region. The interconnection layer extends beyond the transistor and electrically coupled to the transistor at a connection position under the gate structure.

Abstract: An antenna correcting system is provided. The antenna correction system compares amplitudes of antenna signals that are emitted or received respectively by a plurality of antenna units with each other to select one of the amplitudes as target amplitude. The antenna correction system corrects the amplitude of each of the antenna signals according to the target amplitude. As a result, the amplitudes of the antenna signals are the same as each other or approximate to each other. After the amplitudes of the antenna signals are corrected, the antenna correction system compares phases of the antenna signals with each other to select one of the phases as a target phase. The antenna correction system corrects the phase of each of the antenna signals according to the target phase. As a result, the phases of the antenna signals are the same as each other or approximate to each other.

Abstract: An integrated chip including a substrate. A gate layer is over the substrate. A channel layer is over the substrate and vertically spaced apart from the gate layer. A ferroelectric layer is directly between the channel layer and the gate layer. A pair of source/drain electrodes are laterally spaced apart over the channel layer. A plurality of charge traps are along an interface between the ferroelectric layer and the channel layer.

Abstract: An electronic device including a metal casing and at least one antenna module is provided. The metal casing includes at least one window. The at least one antenna module is disposed in the at least one window. The at least one antenna module includes a first radiator and a second radiator. The first radiator includes a feeding end, a first ground end joined to the metal casing, a second ground end, a first portion extending from the feeding end to the first ground end, and a second portion extending from the feeding end to the second ground end. A first coupling gap is between the second radiator and the first portion. A second coupling gap is between at least part of the second radiator and the metal casing, and the second radiator includes a third ground end joined to the metal casing.

Abstract: A separator for a lithium battery and a method for manufacturing the same are provided. The separator includes a substrate layer and a coating layer. The substrate layer is a polyolefin porous film and has a substrate thickness ranging from 10 to 30 micrometers. The coating layer is coated on the substrate layer, and has a coating layer thickness ranging from 1 to 5 micrometers. The coating layer includes a heat-resistant resin material and a plurality of inorganic ceramic particles glued in the heat-resistant resin material. The heat-resistant resin material has a melting point (Tm) or a glass transition temperature (Tg) of not less than 150° C. An average particle size of the inorganic ceramic particles is 10% to 40% of the coating layer thickness of the coating layer. The inorganic ceramic particles are stacked in the coating layer with a height of at least three layers.