Abstract: Provided are improved end-to-end self-supervised pre-training frameworks that leverage a combination of contrastive and masked modeling loss terms. In particular, the present disclosure provides framework that combines contrastive learning and masked modeling, where the former trains the model to discretize input data (e.g., continuous signals such as continuous speech signals) into a finite set of discriminative tokens, and the latter trains the model to learn contextualized representations via solving a masked prediction task consuming the discretized tokens. In contrast to certain existing masked modeling-based pre-training frameworks which rely on an iterative re-clustering and re-training process or other existing frameworks which concatenate two separately trained modules, the proposed framework can enable a model to be optimized in an end-to-end fashion by solving the two self-supervised tasks (the contrastive task and masked modeling) simultaneously.

Abstract: Provided is a projection screen. The projection screen includes a first transparent cover plate, a transparent touch panel, and a first nanoparticle layer, wherein the first nanoparticle layer and the first transparent cover plate are sequentially laminated on one side of the transparent touch panel, and the first nanoparticle layer comprises a plurality of dispersed nanoparticles of different particle sizes. A method for manufacturing a projection screen, a projection display system, and a projection display method are also provided.

Abstract: A method for live streaming recommended content from a content request includes live streaming a first content from a first user side, activating a live reaction at the first user side, and providing the content request at a second user side when the live reaction is enabled. The method also includes determining a second content, sending the content request indicating the second content, and receiving a confirmation at the second user side indicating the content request being sent. The method further includes receiving, at the first user side, the content request sent from the second user side, approving the content request at the first user side, and after the content request is approved, live streaming from the first user side, the second content corresponding to the content request.

Abstract: Embodiments of the present invention disclose a reference signal transmission method, including: sending, by a terminal, a first reference signal and a second reference signal; and correspondingly, receiving, by a network device, the first reference signal and the second reference signal, where the first reference signal is mapped to a plurality of symbols and is used for estimation of channel state information, the second reference signal is mapped to at least two of the plurality of symbols and is used for phase tracking, and a subcarrier to which the second reference signal is mapped on one of the at least two symbols has a same frequency-domain location as a subcarrier to which the second reference signal is mapped on the rest of the at least two symbols. With the foregoing solution, accuracy of channel state information estimation can be improved.

Abstract: The present disclosure relates to a solvent-free moisture curing silicone conformal coating used for the surface protection of circuit boards in the electrical and electronic industries and a preparation method thereof. The solvent-free moisture curing silicone conformal coating comprises (A) a hydrolyzable functional group-terminated linear polysiloxane (I); (B) a hydrolyzable functional group-containing silicone resin with a three-dimensional structure (II); (C) a hydrolyzable functional group-containing crosslinking agent; (D) a catalyst; (E) an adhesion promoter. The solvent-free conformal coating of the present disclosure has low viscosity, high strength and good toughness, and can meet the traditional coating process such as spraying, brushing, dipping, etc., while satisfying the surface protection of the circuit board.

Abstract: A method and device of transparency processing of music. The method comprises: obtaining a characteristic of a music to be played; inputting the characteristic into a transparency probability neural network to obtain a transparency probability of the music to be played; determining a transparency enhancement parameter corresponding to the transparency probability, the transparency enhancement parameter is used to perform transparency adjustment on the music to be played. The present invention constructs a transparency probability neural network in advance based on deep learning and builds a mapping relationship between the transparency probability and the transparency enhancement parameters can be constructed, so that the music to be played can be automatically permeated.

Abstract: A deep learning method-based tonal balancing method, apparatus, and system, the method includes: extracting features from audio data to obtain audio data features, generating audio balancing results by using a trained audio balancing model based on the obtained audio data features. The present invention employs deep neural networks and unsupervised deep learning method to solve the problems of audio balancing of unlabeled music and music of unknown style. The present invention also combines user preferences statistics to achieve a more rational multi-style audio balancing design to meet individual needs.

Abstract: An audio data processing method, an audio data processing apparatus, a terminal, and a computer-readable storage medium are disclosed. In the embodiments of the present invention, a multi-channel audio is processed into left and right channel audio, and the users can experience the effect of surround sound when listening to a target audio.

Abstract: A multi-lens camera module making use of a single image sensor comprising first and second lens modules, an image acquisition module, a refracting module, and a shielding module. The refracting module refracts light from either the first lens module and/or the second lens module, the shielding module allowing light to enter only one of the first or second lens modules by masking the other. The first and second lens modules can share one image acquisition module, which saves cost and interior space of the camera module and thus of the electronic device using such camera module. An electronic device comprising the camera module is also provided.

Abstract: The present invention provides a deep learning based method and system for processing sound quality characteristics. The method comprises: obtaining data characteristics of an audio data to be processed by extracting features from user preference data including the audio data to be processed; based on the data characteristics, generating a sound quality processing result of the audio to be processed by using a trained baseline model; wherein the baseline model is a neural network model trained by using audio data behavioral data, and other relevant data from multiple users or a single user.

Abstract: A biologic sample preparation system that prepares samples for processing includes a frame defining a horizontal plane, a pipette assembly, a sample module and an extraction module. The pipette assembly includes a first pipette. The pipette assembly is movably mounted to the frame in a direction substantially perpendicular to the horizontal plane during operation. The sample module includes a sample plate and is movably mounted to the frame. The sample module is movable substantially parallel to the horizontal plane at least from a sample area spaced from the pipette assembly and a working area proximate the pipette assembly. The extraction module includes an extraction plate and is movably mounted to the frame. The extraction module is movable substantially parallel to the horizontal plane at least from an extraction staging area spaced from the pipette, assembly and the working area proximate the pipette assembly.

Abstract: Provided in the present disclosure are a bifunctional protein which can bind to PD-1 (programmed death receptor-1) and TGF-? (transforming growth factor-?), the medical use of the bifunctional protein, and a preparation method therefor.

Abstract: A dual-lens camera system includes a first lens, a second lens, an image sensor corresponding to a position of the first lens, and a reflecting assembly. The reflecting assembly includes a first reflecting member and a second reflecting member corresponding to a position of the second lens. The first reflecting member is movable between a first position at which an optical path from the first lens to the image sensor is blocked but an optical path from the second reflecting member to the image sensor is not blocked, and a second position at which the optical path from the first lens to the image sensor is not blocked but the optical path from the second reflective member to the image sensor is blocked. Only one image sensor is used in the dual-lens camera system, which saves cost and takes up less internal space of the electronic device.

Abstract: The present invention provides a deep learning based method and system for processing sound quality characteristics. The method comprises: obtaining data characteristics of an audio data to be processed by extracting features from user preference data including the audio data to be processed; based on the data characteristics, generating a sound quality processing result of the audio to be processed by using a trained baseline model; wherein the baseline model is a neural network model trained by using audio data behavioral data, and other relevant data from multiple users or a single user.

Abstract: A gas delivery device includes a nitric oxide generating system. The system has a medium including a source of nitrite ions. A working electrode is in contact with the medium. A Cu(II)-ligand complex is in contact with the working electrode. A reference/counter electrode is, or a reference electrode and a counter electrode are in contact with the medium and separated from the working electrode. An inlet conduit is to deliver nitrogen gas to the medium, and an outlet conduit is to transport a stream of nitrogen gas and nitric oxide from the medium. An inspiratory gas conduit is operatively connected to the outlet conduit to introduce an oxygen-containing gas and form an output gas stream of the gas delivery device.

Abstract: The present invention provides a deep learning based method and system for processing sound quality characteristics. The method comprises: obtaining data characteristics of an audio data to be processed by extracting features from user preference data including the audio data to be processed; based on the data characteristics, generating a sound quality processing result of the audio to be processed by using a trained baseline model; wherein the baseline model is a neural network model trained by using audio data behavioral data, and other relevant data from multiple users or a single user.

Abstract: A grating measuring module includes a light source, a collimating device, a first grating, a second grating, and an image sensing chip. The first grating is arranged on an optical path of the collimating device and the second grating is arranged on an optical path of the first grating. Each of the two gratings comprises a first pattern and a plurality of second patterns locating at the both sides of the first pattern. The first and second gratings can each be attached to an object which may be displaced in relation to another object, so allowing light of a certain pattern to pass depending on the magnitude of the displacement. The image sensing chip arranged on an optical path from the second grating receives light emitted from the light source and forms an image from which a displacement can be calculated and displayed.

Abstract: A laser whose emission is modulated by ultrasound is presented. The laser is usually micron-sized. In response to ultra-sound modulation, the laser emission increases and decreases. Such a change in emission can be detected by external optical detectors. This type of laser can be used as a new type of imaging modality, in which laser emission in combination with sound waves or ultra-sound waves, is used for imaging Laser emission has a much narrower spectral linewidth and stronger intensity than fluorescence and therefore is able to achieve higher sensitivity, whereas sound waves are used to provide a better spatial resolution of the laser emission from the laser. In ultrasound modulated laser based imaging, multiple lasers can be placed inside cells or tissues.

Abstract: A solar energy based mobile EV fast charger system comprising a stationary solar power supply and a mobile EV fast charger installed in a service truck which has a bidirectional Multi-Functional Power Converter System (MFPCS), a solar energy based on-board battery, multiple DC inductors, a alternator power interface and an universal battery interface provides mobile EV charging service for any EV battery.

Abstract: A grating measuring module includes a light source, a collimating device, a first grating, a second grating, and an image sensing chip. The first grating is arranged on an optical path of the collimating device and the second grating is arranged on an optical path of the first grating. Each of the two gratings comprises a first pattern and a plurality of second patterns locating at the both sides of the first pattern. The first and second gratings can each be attached to an object which may be displaced in relation to another object, so allowing light of a certain pattern to pass depending on the magnitude of the displacement. The image sensing chip arranged on an optical path from the second grating receives light emitted from the light source and forms an image from which a displacement can be calculated and displayed.