Abstract: A training technique trains a neural network having sparsely-activated sub-networks. It does so by processing plural batches of training data in two respective passes of the neural network, yielding first prediction information and second prediction information. For each batch, the technique randomly assigns different sub-networks in the first and second passes of the neural network to process the batch. Over the course of training, the technique attempts to minimize loss information, which describes the difference between the first prediction information and ground-truth information, and the difference between the second prediction information and the ground-truth information. Simultaneously, the technique attempts to minimize divergence information, which describes the divergence of the first prediction information from the second prediction information (and vice versa).

Abstract: An electrode includes a microporous titanium substrate coated with a catalytic layer, and the catalytic layer includes magnetic SnO2—Sb particles. The magnetic SnO2—Sb particles are attached to the microporous titanium substrate through an external magnetic field. The microporous titanium substrate includes a plurality of membrane pores having a pore size of 5-50 ?m that is smaller than a particle size of the magnetic SnO2—Sb particles.

Abstract: A conjugate compounds or pharmaceutically acceptable salt thereof, comprises a payload and two or more kinds of cell-interacting molecules. The cell-interacting molecules are ligands capable of specifically binding to a cell surface receptor. A method of treating diseases, comprises delivering a payload to a subject.

Abstract: A method, system, and device for planning a path for a forced landing of an aircraft based on image recognition are provided. The method includes: calculating an endurance distance of an aircraft based on sensor data and meteorological information; obtaining an alternative landing area by a satellite image containing contour information and a terrain image recognition model; obtaining a current satellite image of the alternative landing area and determining a landing area; and selecting a landing site by a landing site decision model and generating a path for a forced landing, such that the aircraft completes a forced landing task according to the path for the forced landing. The method, system, and device can automatically recognize image information, select a best landing site, and generate a path for a forced landing to assist a pilot in performing a forced landing task.

Abstract: Video playback quality detection methods, systems, and apparatuses are described. The method can include obtaining a first video data packet of a target video from a bearer network and obtaining a data byte quantity of first video data included in the first video data packet. The method may also include determining a buffer byte quantity stored in a video quality detection buffer. Furthermore, the method can include obtaining video frame data rate information corresponding to the target video and reducing the buffer byte quantity stored in the video quality detection buffer based on a video frame rate of the target video and the video frame data rate information of the target video. The method may further include determining a playback status of playing the target video on a video player, and determining video playback quality of the target video based on the playback status.

Abstract: Systems and methods of adjusting zone associated risks of a coverage zone covered by one or more sensors of an autonomous driving vehicle (ADV) operating in real-time are disclosed. As an example, the method includes defining a performance limit detection window associated with a first sensor based on a mean time between failure (MTBF) lower limit of the first sensor and a MTBF upper limit of the first sensor. The method further includes determining whether an operating time of the ADV operating in autonomous driving (AD) mode is within the performance limit detection window associated with the first sensor. The method further includes in response to determining that the operating time of the ADV operating in AD mode is within the performance limit detection window of the first sensor, adjusting a zone associated risk of the coverage zone to a performance risk of a second sensor.

Abstract: A vehicle control unit (e.g., a control unit for an automobile) receives feedback from an intelligent voltage/current sensor and a DC/DC controller. The DC/DC controller comprises a first switch for controlling power from a primary power source (e.g., low voltage power supplied from a high voltage battery). The intelligent voltage/current sensor senses power output from the primary power source. The vehicle control unit processes feedback from the intelligent voltage/current sensor and/or the DC/DC controller to determine if a failure has occurred in the primary power source. In response to determining the failure in the primary power source, the vehicle control unit disables the power from the primary power source using a second switch (e.g., a switch in a relay).

Abstract: The present invention relates to the technical field of wastewater treatment, and discloses a bioaugmentation treatment process for lithium battery producing wastewater. The method comprises the following steps: 1) introducing wastewater into a hydrolytic acidification tank, and adding Enterobacter sp. NJUST50 and activated sludge to the hydrolytic acidification tank for hydrolytic acidification treatment; 2) introducing the effluent into an anoxic tank, and adding Enterobacter sp. NJUST50 and anaerobic activated sludge for anoxic treatment; 3) introducing the effluent into an aerobic tank, and adding Enterobacter sp. NJUST50 and aerobic activated sludge for aerobic treatment; 4) introducing the effluent into an anoxic filter tank, and adding Enterobacter sp. NJUST50 and anaerobic activated sludge to the filter tank for treatment; and 5) introducing the effluent into a biological aerated filter tank, and adding a sludge mixture of Enterobacter sp.

Abstract: The present invention relates to methods for producing an induced pluripotent stem (iPSC), the method comprising the following steps in order: culturing a somatic cell in a first culture condition adapted to promote the reprogramming of the cell towards a pluripotent state; culturing the cell in a second culture condition adapted to promote a hypomethylated DNA state in the cell; and culturing the cell in a third culture condition adapted to promote a primed pluripotent state, thereby producing an iPSC from a somatic cell. The invention also relates to cells and compositions resulting from those methods.

Abstract: Systems and methods are provided for facilitating the building and use of natural language understanding models. The systems and methods identify a plurality of tokens and use them to generate one or more pre-trained natural language models using a transformer. The transformer disentangles the content embedding and positional embedding in the computation of its attention matrix. Systems and methods are also provided to facilitate self-training of the pre-trained natural language model by utilizing multi-step decoding to better reconstruct masked tokens and improve pre-training convergence.

Abstract: The present invention provides a method for reprogramming a human somatic cell to a cell exhibiting at least one characteristic of a trophoblast stem cell (TSC), the method comprising the following steps in order: a) increasing the protein expression of one or more factors in the somatic cell, wherein the factors are for reprogramming the somatic cell towards a pluripotent state; b) culturing the cell for a sufficient time and under conditions to allow the reprogramming of the cell towards a pluripotent state; c) contacting the cell with a culture medium suitable for sustaining trophoblast stem cells (TSC); and d) culturing the cell in the TSC medium for a sufficient time and under conditions to allow the cell to exhibit at least one characteristic of a TSC, thereby reprogramming the somatic cell to a cell exhibiting at least one characteristic of a TSC.

Abstract: An apparatus is described. The apparatus includes an accelerator to be coupled to a memory region that the accelerator shares with a virtualization environment comprising a guest OS, a guest VM and an SSD device driver. The accelerator is to forward a submission queue doorbell setting made by the SSD device driver in the shared memory to a corresponding submission queue doorbell in an SSD controller.

Abstract: A method for producing a stainless steel foil comprises the following steps: rolling a raw material steel coil in an intermediate rolling process comprising three rolling processes; rolling the finished product in the rolling process; annealing and straightening the finished product. The application also provides a stainless steel foil prepared by the method. The method can solve the defect that the high strength and high plasticity of the ultra-thin stainless steel foil are difficult to meet at the same time.

Abstract: A configuration apparatus includes a processor and a memory. The processor is configured to obtain first data of a first device, the first data indicating that the first device is ready for network configuration. The memory is configured to store network configuration data of a second device. The processor is further configured to: obtain the network configuration data of the second device, and send the network configuration data to the first device, the network configuration data being used to configure the first device to be in a same network connection state as the second device. A power path exists between the first device and the second device, and at least one of the first data or the network configuration data is transmitted through the power path.

Abstract: The present invention relates to the technical field of electrocatalytic electrode preparation, and discloses a lead dioxide-carbon nanotube adsorptive electrochemical submicroelectrode, a preparation method, and use thereof. The electrochemical submicroelectrode according to the present invention comprises multiple layers of orderly arranged spherical lead dioxide submicroholes communicating with each other, where the carbon nanotubes are partially or completely inserted (in the form of twigs) in the lead dioxide hole and in the wall of the hole. The combined effect of adsorption and catalysis inside the submicroreactor effectively solves the problems of low catalytic efficiency and diffusion control associated with the conventional flat lead dioxide electrode, thus greatly improving the electrochemical catalytic performance of the electrode.

Abstract: The present invention provides a compound of Formula (I) or a pharmaceutically acceptable salt thereof; a method for manufacturing the compounds of the invention, and its therapeutic uses. The present invention further provides a combination of pharmacologically active agents and a pharmaceutical composition.

Abstract: The present invention belongs to the bioengineering field, and relates to a method for fermentation production of L-theanine by using an Escherichia coli genetically engineered bacterium. The engineered bacterium is obtained by serving a strain as an original strain, wherein the strain is obtained after performing a single copy of T7RNAP, a dual copy of gmas, xylR knockout, and sucCD knockout on an Escherichia coli W3110 genome, and by integrating genes xfp, pta, acs, gltA, and ppc, and knocking out ackA on the genome. The present invention has a high yield, and stable production performance; after 20-25 h, L-theanine has a titer of 75-80 g/L, and the yield is up to 52-55%. The fermentation broth is purified by membrane separation in combination with a cation-anion resin series technique. Moreover, the one-step crystallization yield is 72.3% and the L-theanine final product has a purity of 99%.