Abstract: A system and method for video temporal action proposal generation are provided. It processes video features extracted from the input video through an encoder to obtain video encoding features with global information, extracts corresponding interest segment features from the video encoding features using pre-trained proposal segments, and provides them to the decoder. The decoder generates segment features based on the interest segment features corresponding to each proposal segment and the pre-trained proposal features. These are then provided to the prediction module, generating temporal action proposal results based on the decoder's segment features. The solution in embodiments of the present invention can effectively capture global context information of the video, obtaining video encoding features with stronger representational capabilities.

Abstract: The embodiments of the invention relate to a workflow creation method, apparatus, device, and a storage medium. One method provided herein includes: in response to receiving an editing operation, adding a first node associated with a canvas configuration to a node connected graph; determining canvas configuration information based on a set of configuration controls associated with the first node, the canvas configuration information indicating at least a layout of a set of elements in the canvas, the set of elements including a first element configured by referencing a second node associated with the first node; and creating a target workflow based on the node connected graph, the first node in the target workflow being configured to output an image processing result based on the canvas configuration information. In this way, the embodiments of the present disclosure can improve the efficiency of image processing using a workflow.

Abstract: The present invention relates to a method for authenticating a document comprising at least one line of anti-counterfeit patterns spaced from each other, the positions of the anti-counterfeit patterns being random, and a partial anti-counterfeit pattern being provided at an edge line of said document, the method comprising the following steps of: receiving a digital image of said document, said digital image comprising said partial anti-counterfeit pattern in a first edge area corresponding to said edge line; selecting said first edge area; copy-pasting said first edge area adjacent to a second edge area of the digital image to generate a combined digital image, said second edge area being opposite to said first edge area, wherein content, which is comprised in said second edge area and located on the same line as said partial anti-counterfeit pattern, and said partial anti-counterfeit pattern jointly form a combined pattern; authenticating said document by taking into account said combined pattern.

Abstract: Use of an inducible extracellular vesicle as a medicament for skin, health care products or skin care products. A composition, comprising the inducible vesicle and the medicaments or care products for skin.

Abstract: A glass composite includes a first glass member and a second glass member. The first glass member and the second glass member are at least partially connected with each other at the surfaces; and the glass composite has a light transmittance not lower than 95% of the light transmittance of the one, with the lower light transmittance, of the first glass member and the second glass member.

Abstract: An electronic device housing, a manufacturing method thereof, and an electronic device are provided. The housing includes a bottom plate, and a middle frame connected to the bottom plate. The middle frame is made of glass, and the bottom plate is made of sapphire; or the bottom plate is made of glass, and the middle frame is made of sapphire. An interface-free continuous connection is provided between the bottom plate and the middle frame.

Abstract: A ceramic-like structure includes: a transparent substrate; an ink layer; and a ceramic-like film layer. The ceramic-like film layer is disposed between the transparent substrate and the ink layer, and includes a first high-refractive-index material layer, a first low-refractive-index material layer, a second high-refractive-index material layer, and a second low-refractive-index material layer that are stacked on each other. Refractive indexes of the first high-refractive-index material layer and the second high-refractive-index material layer are greater than about 2.3, and refractive indexes of the first low-refractive-index material layer and the second low-refractive-index material layer are less than about 1.6.

Abstract: Approaches are generally described for corrupted speech detection in voice-based computer interfaces. First input data including first audio data representing a user utterance may be received. First data representing the first audio data may be generated using a first encoder. First text data representing a transcription of the user utterance may be generated. Second data representing the first text data may be generated using a second encoder different from the first encoder. Third data may be generated by combining the first data and the second data. The third data may be sent to a classifier network trained to predict a relevant corruption state for speech processing inputs. The classifier network may determine that the first input data corresponds to a first corruption state.

Abstract: Disclosed herein is a sphingan oligosaccharide composition, its uses, and a process for preparing the same.

Abstract: A system and method for video temporal action proposal generation are provided. It processes video features extracted from the input video through an encoder to obtain video encoding features with global information, extracts corresponding interest segment features from the video encoding features using pre-trained proposal segments, and provides them to the decoder. The decoder generates segment features based on the interest segment features corresponding to each proposal segment and the pre-trained proposal features. These are then provided to the prediction module, generating temporal action proposal results based on the decoder's segment features.

Abstract: The present disclosure provides a gradient glass and a preparation method and use thereof. The gradient glass includes: a glass body, a haze of at least a part of an outer surface of the glass body being greater than 0, the haze gradually changing in a predetermined direction, and a light transmittance at any position of the glass body being greater than 85%.

Abstract: A decorative plate and a manufacturing method therefor, and an electronic device are provided. The decorative plate includes a substrate and a decorative layer arranged on a surface of the substrate. A surface of a side of the decorative layer distant from the substrate includes at least one unit block; in a first direction, the unit block includes a plurality of texture units which are continuously arranged; each texture unit includes a plurality of texture line grooves, and an extension direction of the plurality of texture line grooves is the same; an included angle between the extension direction of the plurality of texture line grooves and the first direction is an extension angle; in the first direction, extension angles of the plurality of texture line grooves in different texture units gradually change.

Abstract: A method for forming a glass composite includes: providing a first glass member and a second glass member; activating at least a part of a surface of the first glass member to form a first activated surface, with unsaturated chemical bonds formed on the first activated surface of the first glass member; activating at least a part of a surface of the second glass member to form a second activated surface, with unsaturated chemical bonds on the second activated surface of the second glass member; and connecting at least partially the first glass member and the second glass member with each other on a contact interface at a contacting position of the first glass member and the second glass member to form the glass composite.

Abstract: A glass device includes a glass body. The glass body has a first surface. The first surface includes a plurality of first frosted regions distributed spaced apart from each other, two adjacent first frosted regions are connected by a second frosted region, and each of the first frosted regions and the second frosted region have a shape of a straight strip. Each of the first frosted regions has a plurality of first protruding structures, the second frosted region has a plurality of second protruding structures, and a height and a length of one of the second protruding structures are less than a height and a length of one of the first protruding structures, respectively.

Abstract: An electronic device housing, an electronic device and a compound body are provided. The electronic device housing comprises a frame; a sealing layer, disposed on at least a part of an outer surface of the frame, and including a plurality of sub-sealing layers laminated in sequence; and a back case, attached to the frame by the sealing layer, wherein two adjacent sub-sealing layers have different compositions.

Abstract: A laminated glass includes at least two glass members and at least one adhesive film disposed in a stacked manner, where the glass members and the adhesive film are alternately disposed; and decorative layers provided on surfaces of the at least two glass members facing toward the at least one adhesive film. At least two of the decorative layers independently comprise an etched texture, an optical coating layer, and a pattern layer sequentially stacked, or at least two of the decorative layers independently comprise the etched texture, the pattern layer, and the optical coating layer sequentially stacked.

Abstract: A glass composite includes a first glass member and a second glass member. The first glass member and the second glass member are at least partially connected with each other at the surfaces; and the glass composite has a light transmittance not lower than 95% of the light transmittance of the one, with the lower light transmittance, of the first glass member and the second glass member.

Abstract: A glass composite includes a first glass member and a second glass member. The first glass member and the second glass member are at least partially connected with each other at the surfaces. The glass composite has a light transmittance not lower than 95% of the light transmittance of the one, with the lower light transmittance, of the first glass member and the second glass member.

Abstract: Features are disclosed for training and using a word embedding model configured to receive textual and context data associated with an utterance of a user. A word embedding model can be trained with text data and context data to account for context associated with the text data. The word embedding model can receive an input vector including text data and one or more sets of context data associated with the text data and perform word embedding based on the input vector. In some embodiments, the input vector can include an automatic speech recognition (“ASR”) confidence score generated by an ASR model and one or more labels generated by an NLU model. In some embodiments, the input vector can include user information associated with the user.

Abstract: Disclosed herein is a sphingan oligosaccharide composition, its uses, and a process for preparing the same.