Abstract: The present invention discloses an environmentally friendly emission system for a kitchen waste incinerator, comprising a waste incinerator, a dry desulfurization tower, a dust collector, and a chimney; a flue gas emission pipe of the waste incinerator is connected to the dry desulfurization tower; an outlet of the dry desulfurization tower is connected to an inlet of the dust collector; an outlet of the dust collector is connected to the chimney; a high-temperature heat exchanger and a low-temperature heat exchanger are sequentially provided on the flue gas pipeline between the waste incinerator and the dry desulfurization tower; a branch recirculation pipeline is provided on the pipeline between the high-temperature heat exchanger and the low-temperature heat exchanger and is connected to an inlet of a recirculation fan; the outlet of the recirculation fan are respectively connected to each air chamber provided on the waste incinerator through pipeline.

Abstract: A specialized air distribution system for a waste incinerator with sludge co-incineration includes a furnace, primary air circulation cycle and secondary air circulation cycle. The primary air circulation cycle involves arranging a flue gas suction port on side wall of upper part of the furnace connected to a first ejector, connecting a primary air fan to the first ejector, connecting an outlet of first ejector to a primary air main pipe, and connecting the primary air main pipe to multiple primary air inlets at lower part of the furnace through multiple output ports. The secondary air circulation cycle involves connecting a steam drum arranged above the furnace to a second ejector via a high-pressure steam pipe, connecting an outlet of the second ejector to a secondary air inlet on side wall of the rear arch outlet of the furnace, and connecting a secondary air fan to the second ejector.

Abstract: System, method and/or computer readable medium for audio visual sound source Separation that applies cross-modal meta consistency learning. Inputs include an audio spectrogram that represents first sounds second sounds, a first video of a first sound producing object, and a second video of a second sound producing object. Audio features and audio tokens are generated by applying audio encoders to the audio spectrogram. Visual tokens are generated by applying a visual encoder to the first video and the second video. Respective audio-visual features are obtained based on combining the audio token with the respective visual tokens. Based on the respective audio-visual features, first and second separated audio masks are generated by applying a decoder.

Abstract: A power on/off control circuit and an electronic device is provided, the circuit including: a switching element, a first semiconductor switching circuit including a first semiconductor switching element, a second semiconductor switching circuit including a second semiconductor switching element, and a holding circuit connected between a first terminal of the first semiconductor switching element and a second terminal of the second semiconductor switching element. When the switching element is closed, the first semiconductor switching element is turned on. When the first semiconductor switching element is turned on, the second semiconductor switching element is turned on, and a power on signal is output via the second terminal of the second semiconductor switching element. When the second semiconductor switching element is turned on, the holding circuit is operable to maintain the on state of the first semiconductor switching element. The control circuit can stably lock a started state.

Abstract: A low-power electronic pipette has a charging management module (10), a control module (20), a motor driving module (30), a display module (40), a communication module (50), a sensing module (60), and a battery module (70). The charging management module, the communication module, and the sensing module are each in two-way communication with the control module. The display module and the motor driving module are each in one-way communication with the control module. The battery module supplies energy to the low-power electronic pipette. When the sensing module acquires a signal and determines that the low-power electronic pipette has not been operated for a predetermined period and is not in a charging state, the low-power electronic pipette enters a low-power mode, in which the control module turns off the charging management module, the motor driving module, the display module, and the communication module, and the control module enters a sleep state.

Abstract: Methods and systems that train a gaze estimation model are provided. The methods and systems generating a mirror image from a given image, generating, by a neural network implementing the model, a first gaze prediction based the given image and a mirror gaze prediction based on the mirror image, generating a symmetry loss value based on a comparison between the first gaze prediction and the mirror gaze prediction, generating, by the neural network, a reconstructed image based on at least one of the given image and the mirror image, generating a reconstruction loss value based on a comparison between the reconstructed image and the at least one of the given image and the mirror image, and updating at least some parameters of the neural network based on a combination of the symmetry loss value and the reconstruction loss value.

Abstract: This disclosure relates to the field of image processing, and discloses an apparatus, method and readable storage medium for image processing model training. The model includes multiple prompt modules and multiple image processing modules corresponding to multiple prompt modules. The method comprising: input N sample images into multiple prompt modules to add corresponding prompt information to the N sample images. Thus, N prompt sample images corresponding to each image processing module may be obtained. And then input the N prompt sample images into each image processing module, to obtain N prediction results corresponding to the each image processing module. Then parameters of multiple image processing modules and multiple prompt modules can be adjusted based on N prediction results and N reference results corresponding to each image processing module.

Abstract: Methods and devices for gaze-based device interactions are described. During display of a notification on a touch-sensitive display of an electronic device, gaze tracking is performed to determine point of gaze locations. It is determined, within a defined time period, that recent point of gaze locations fall within a proxy region of the notification. In response to determining that recent point of gaze locations fall within the proxy region of the notification, the electronic device is operated in a one-handed mode. In the one-handed mode, a touch input detected on the touch-sensitive display outside the notification is processed as an interaction with the notification.

Abstract: The present invention discloses a Cpf1 protein and its use in gene editing, which relates to the field of biotechnology. The present invention provides a Cas12a nuclease named as the Cas12a-68 protease, which has a wider PAM recognition range than traditional nucleases and is capable of recognizing targets of various genes. When editing nucleic acid fragments, using the new nuclease of the present invention is capable of overcoming the shortcomings of the traditional Cas12a and increasing the PAM recognition range. These advantages make the present invention more suitable for editing a wider range of nucleic acid fragments and lay a foundation for gene editing applications.

Abstract: There is provided a method, apparatus and system for adapting a machine learning model for optical flow prediction. A machine learning model can be trained or adapted based on compressed video data, using motion vector information extracted from the compressed video data as ground-truth information for use in adapting the model to a motion vector prediction task. The model so adapted can accordingly be adapted for the similar task of optical flow prediction. Thus, the model can be adapted at test time to image data which is taken from an appropriate distribution. A meta-learning process can be performed prior to such model adaptation to potentially improve the model's performance.

Abstract: A graph calculation method of RNA similarity analysis, an apparatus, a device, and a medium are provided. The method includes: converting sequence data of a looked-up RNA into a looked-up RNA structure graph; obtain a first similarity between the looked-up RNA structure graph and a target RNA structure graph; obtaining a second similarity based on the number of base constituent structures in the looked-up RNA structure graph and the number of base constituent structures in the target RNA structure graph; reconstructing the looked-up RNA structure graph based on the base constituent structures in the looked-up RNA structure graph to generate a looked-up RNA higher-order graph; and analyzing similarity between the looked-up RNA higher-order graph and a target RNA higher-order graph to obtain a third similarity; and obtaining a final similarity between the looked-up RNA and the target RNA based on the first similarity, the second similarity, and the third similarity.

Abstract: A semiconductor structure may include a substrate having a surface, an isolation structure formed on the surface, an active region formed on the surface adjacent to the isolation structure, a gate extended over the isolation structure and the active region, and a source region formed within the active region. The source region may include a first subregion formed adjacent to a first portion of the gate and the isolation structure, the first subregion having a first doping and a second subregion formed adjacent to a second portion of the gate, wherein the second subregion has a second doping different from the first doping.

Abstract: A semiconductor structure may include a substrate, an active region formed on the substrate, a gate region extended over the active region along a first direction parallel to a surface of the substrate, and a source region formed within the active region and adjacent to the gate region in a second direction parallel to the surface and perpendicular to the first direction. The source region may comprise a first subregion and a second subregion such that the second subregion is between the first subregion and the gate region. The first subregion may comprise a first width along the first direction and the second subregion may comprise a second width along the first direction. The first width may be less than the second width.

Abstract: The present invention discloses a Cpf 1 protein and its use in gene editing, which relates to the field of biotechnology. The present invention provides a novel Cas12a nuclease named as the Cas12a-68 protease, which has a wider PAM recognition range than traditional nucleases and is capable of recognizing targets of various genes. When editing nucleic acid fragments, using the new nuclease of the present invention is capable of overcoming the shortcomings of the traditional Cas12a and increasing the PAM recognition range. These advantages make the present invention more suitable for editing a wider range of nucleic acid fragments and lay a foundation for gene editing applications.

Abstract: Methods, devices and computer-readable media for processing a compressed video to perform an inference task are disclosed. Processing the compressed video may include selecting a subset of frame encodings of the compressed video, or zero or more modalities (RGB, motion vectors, residuals) of a frame encoding, for further processing to perform the inference task. Pre-existing motion vector and/or residual information in frame encodings of the compressed video are leveraged to adaptively and efficiently perform the inference task. In some embodiments, the inference task is an action recognition task, such as a human action recognition task.

Abstract: A power on/off control circuit and an electronic device is provided, the circuit including: a switching element, a first semiconductor switching circuit including a first semiconductor switching element, a second semiconductor switching circuit including a second semiconductor switching element, and a holding circuit connected between a first terminal of the first semiconductor switching element and a second terminal of the second semiconductor switching element. When the switching element is closed, the first semiconductor switching element is turned on. When the first semiconductor switching element is turned on, the second semiconductor switching element is turned on, and a power on signal is output via the second terminal of the second semiconductor switching element. When the second semiconductor switching element is turned on, the holding circuit is operable to maintain the on state of the first semiconductor switching element. The control circuit can stably lock a started state.

Abstract: Systems, methods and computer-readable medium for predicting a depth for a video frame are disclosed. An example method may include steps of: receiving a plurality of training data, each comprising a set of consecutive video frames and a depth representation of a subsequent video frame to the consecutive video frames; receiving a pre-trained neural network model f? having a plurality of weights ?; while the pre-trained neural network model f? has not converged: computing a plurality of second weights ?i?, based on each set of consecutive video frames, and updating the plurality of weights ?, based on the plurality of training data and the plurality of second weights ?i?; receiving a plurality of new consecutive video frames with consecutive timestamps; and predicting a depth representation of video frame immediately subsequent to the new consecutive video frames based on the updated plurality of weights ?.

Abstract: Methods and modules for accelerating inference computations in transformer models using edge devices includes partitioning inputs for each layer and synchronizing between transformer layers. A method includes receiving a transformer input, partitioning the transformer input into two or more first-stage divisions, processing each first-stage division into a processed first-stage division, and combining the processed first-stage divisions into a first output. A module includes a computing device for partitioning a transformer input into two or more divisions, transmitting each of the divisions, and receiving processed divisions, as well as two or more transformer processing units, each for receiving a division from the computing device, processing the division into a processed division, and sending the processed division to the computing device.

Abstract: System, method and/or computer readable medium for audio visual sound source Separation that applies cross-modal meta consistency learning. Inputs include an audio spectrogram that represents first sounds second sounds, a first video of a first sound producing object, and a second video of a second sound producing object. Audio features and audio tokens are generated by applying audio encoders to the audio spectrogram. Visual tokens are generated by applying a visual encoder to the first video and the second video. Respective audio-visual features are obtained based on combining the audio token with the respective visual tokens. Based on the respective audio-visual features, first and second separated audio masks are generated by applying a decoder.

Abstract: This disclosure provides for methods and system for meta few-shot class incremental learning. According to an aspect a method is provided. The method includes obtaining at least one weight attention map of a first network and updating weights of a second network using the at least one weight attention map, where the second network is a modulatory network. The method further includes generating at least one feature attention map of the second network based on the at least one weight attention map of the first network and a set of input images of at least one class. The method further includes generating at least one feature map of the first network based on the set of input images of the at least one class, and updating the at least one feature map of the first network based on the feature attention map of the second network.