Abstract: A motor cooling structure, including: branch flow-channels, shell flow-channels, end cover flow-channels, a liquid inlet and a liquid outlet. A plurality of the branch flow-channels are circumferentially arranged on a stator of a motor around an axis of the motor. The shell flow-channels includes a liquid inlet flow-channel, a plurality of shell long flow-channels and a liquid outlet flow-channel, which are circumferentially arranged on a reducer shell around the axis of the motor. The end cover flow-channels includes a plurality of the end-cover long flow-channels are circumferentially arranged on a motor end cover around the axis of the motor. The shell flow-channels, the plurality of branch flow-channels and the end cover flow-channels form a continuous total flow-channel. The liquid inlet and liquid outlet both disposed on the reducer shell are in communication with the liquid inlet flow-channel and liquid outlet flow-channel, respectively.

Abstract: To present one or more content items to users of an online system, the online system identifies a content evaluation pipeline including an order of a plurality of stages having one or more computer models for evaluating a likelihood of user interaction with a content item. The content evaluation pipeline selects a decreasing number of content items, from each stage of the order, according to the order of the stages. The online system optimizes the selection of content items selected at the plurality of stages of the content evaluation pipeline by training the computer models to predict content selection values that the subsequent model would generate for a content items in a training data set and content items that the subsequent model would select for input to the next stage of the content evaluation pipeline.

Abstract: In certain aspects, a power amplification device, a power amplification system, and an operation method thereof are disclosed. The power amplification system includes a power amplification device and a control unit. The power amplification device includes a plurality of power amplification circuits, a power supply network configured to supply a power source from a plurality of power sources to the plurality of power amplification circuits, and a transformer network configured to provide an output power that matches a transmit power level. The control unit includes a memory storing code and a processor coupled to the memory. When the code is executed, the processor is configured to determine a plurality of circuit configurations for the plurality of power amplification circuits based on the transmit power level, and configure the plurality of power amplification circuits using the plurality of circuit configurations, respectively, so that the output power matches the transmit power level.

Abstract: An electric drive module and an electric drive equipment are provided, the electric drive module comprises a housing, a force output assembly, a flexible gear, a rotor, a stator, a wave generator, and a cooling pipe, the stator is configured to drive the rotor to rotate relative to the housing, when the rotor rotates, the wave generator drives the flexible gear to deform to drive the rigid gear to rotate, at least part of the cooling pipe is received in and closes to the stator. The electric drive module is compact and space saving, the flexible gear is secured to the housing, the rotation of the rigid gear driven by the deformation of the flexible outputs power, which is low rotational inertia and decreases vibration, the cooling pipe arranged in the stator can directly dissipate heat from the stator with high heat dissipation efficiency.

Abstract: For a content item with unknown tasks performed by a viewing user on an online system, the online system predicts a likelihood of interacting with each content item using a prediction model associated with a plurality of tasks. The prediction model comprises a plurality of independent layers, a plurality of shared layers and a plurality of separate layers. Each independent layer is configured to extract features, for each task, that are not shared across the plurality of tasks. The plurality of shared layers are configured to extract common features that are shared across the plurality of tasks. Each separate layer is configured to predict likelihood of the viewing user performing a task associated with the separate layer based on the features extracted from the plurality of independent layers and the plurality of shared layers.

Abstract: Systems and methods for controlling air pollution using a venturi scrubber with a venturi throat having multiple variable apertures. In some embodiments, the venturi throat comprises a first plate and a second plate, each with multiple apertures. The plates may move relative to each other. The dimensions of the apertures may be determined based on the movement of two plates.

Abstract: A method for controlling heating of a battery pack is provided, in which, the method includes: acquiring a temperature of the battery pack; and controlling a motor controller to output a current to a motor, if the temperature of the battery pack is lower than a preset temperature threshold, so as to generate a magnetic field in the motor having a magnetic pole direction consistent with or opposite to a magnetic pole direction of a motor rotor and to enable the motor to maintain a stationary state.

Abstract: A motor cooling structure, a drive assembly and a vehicle. The motor cooling structure includes: branch flow-channels (100), shell flow-channels (200), end cover flow-channels (300), a liquid inlet (201) and a liquid outlet (202). A plurality of the branch flow-channels (100) are circumferentially arranged on a stator (1) of a motor around an axis of the motor. The shell flow-channels (200) includes a liquid inlet flow-channel (211), shell long flow-channels and a liquid outlet flow-channel (212). The liquid inlet flow-channel (211), the plurality of shell long flow-channels and the liquid outlet flow-channel (212) are circumferentially arranged on a reducer shell (2) around the axis of the motor. The end cover flow-channels (300) includes end-cover long flow-channels, and a plurality of the end-cover long flow-channels are circumferentially arranged on a motor end cover (3) around the axis of the motor.

Abstract: To select the content to be presented to the user, a first latent vector is determined for a content item based on a first object associated with the content item. A second latent vector is determined for the content item based on a second object associated with the content item. A content item vector is then determined based on the first and second latent vectors. Furthermore, a user vector is determined based on interactions of the user with the first set of content objects and the second set of content objects. A score indicative of the likelihood of the user interacting with the content item is determined based on the content item vector and the user vector.

Abstract: A method of charging a battery and a system using the same are provided. The method includes: detecting a terminal voltage of a battery; When the terminal voltage is less than a voltage threshold, performing quick charging which includes: calculating a rising rate of the terminal voltage, comparing the rising rate with a rising rate reference value, and controlling the terminal voltage according to the comparison result; and when the terminal voltage is not less than the voltage threshold, performing low current floating charging which includes: intermittently charging the battery by a predetermined floating charging period, calculating a falling rate of the terminal voltage, comparing the falling rate with a falling rate reference value, controlling the terminal voltage according to the comparison result, in a first time duration of the floating charging period, and suspending charging in a second time duration of the floating charging period.

Abstract: A preprocessing module of a neural network has a first input and second input. The module generates multiple, different first latent vector representations of its first input, and multiple, different second latent vector representations of its second input. The module then models pairwise interactions between every unique pairwise combination of the first and second latent vector representations. The module then produces an intermediate output by combining the results of the modeled pairwise interactions.

Abstract: In one embodiment, a method includes a preprocessing stage of a neural network model, where the preprocessing stage includes first and second preprocessing modules. Each of the two modules has first input that may receive a dense input and a second input that may receive a sparse input. Each module generates latent vector representations of their respective first and second inputs, and combine the latent vectors with the original first input to define an intermediate output. The intermediate output of the first module is fed into the first input of the second module.

Abstract: Systems and methods for controlling air pollution using a venturi scrubber with a venturi throat having multiple variable apertures. In some embodiments, the venturi throat comprises a first plate and a second plate, each with multiple apertures. The plates may move relative to each other. The dimensions of the apertures may be determined based on the movement of two plates.

Abstract: For a content item with unknown tasks performed by a viewing user on an online system, the online system predicts a likelihood of interacting with each content item using a prediction model associated with a plurality of tasks. The prediction model comprises a plurality of independent layers, a plurality of shared layers and a plurality of separate layers. Each independent layer is configured to extract features, for each task, that are not shared across the plurality of tasks. The plurality of shared layers are configured to extract common features that are shared across the plurality of tasks. Each separate layer is configured to predict likelihood of the viewing user performing a task associated with the separate layer based on the features extracted from the plurality of independent layers and the plurality of shared layers.

Abstract: An optimized computer architecture for training an neural network includes a system having multiple GPUs. The neural network may be divided into separate portions, and a different portion is assigned to each of the multiple GPUs. Within each GPU, its portion is further divided across multiple training worker threads in multiple processing cores, and each processing core has lock-free access to a local parameter memory. The local parameter memory of each GPU is separately, and individually, synchronized with a remote master parameter memory by lock memory access. Each GPU has a separate set of communication worker threads dedicated to data transfer between the GPU and the remote parameter memory so that the GPU's training worker threads are not involved with cross GPU communications.

Abstract: For a content item with unknown tasks performed by a viewing user on an online system, the online system receives a plurality of content items associated with a viewing user. The online system derives a feature vector for each content item. The online system predicts a likelihood of interacting with each content item using a prediction model associated with a plurality of tasks. The prediction model comprises a plurality of shared layers and a plurality of separate layers. The plurality of shared layers are configured to extract common features that are shared across the plurality of tasks. Each separate layer is configured to predict likelihood of the viewing user performing a task associated with the separate layer based on the common features. The online system scores each content item based on predicted likelihood of each task. The online system ranks the plurality of content items based on the scoring.

Abstract: The present invention discloses a power supply circuit and a UPS auxiliary power supply system having the same. The power supply circuit comprises a transformer, a primary-side switch unit and a secondary-side switch unit. When the input power supply is powered on, the second switch of the secondary-side switch unit maintains in an off state, the first switch of the primary-side switch unit performs an on-off action, and the input power supply supplies power to the load and the energy storage unit. When the input power supply is powered down, the first switch of the primary-side switch unit maintains in an off state, the second switch of the secondary-side switch unit performs an on-off action, and the energy storage unit supplies power to the load.

Abstract: A method of charging a battery and a system using the same are provided. The method includes: detecting a terminal voltage of a battery; When the terminal voltage is less than a voltage threshold, performing quick charging which includes: calculating a rising rate of the terminal voltage, comparing the rising rate with a rising rate reference value, and controlling the terminal voltage according to the comparison result; and when the terminal voltage is not less than the voltage threshold, performing low current floating charging which includes: intermittently charging the battery by a predetermined floating charging period, calculating a falling rate of the terminal voltage, comparing the falling rate with a falling rate reference value, controlling the terminal voltage according to the comparison result, in a first time duration of the floating charging period, and suspending charging in a second time duration of the floating charging period.

Abstract: The present invention discloses a power supply circuit and a UPS auxiliary power supply system having the same. The power supply circuit comprises a transformer, a primary-side switch unit and a secondary-side switch unit. When the input power supply is powered on, the second switch of the secondary-side switch unit maintains in an off state, the first switch of the primary-side switch unit performs an on-off action, and the input power supply supplies power to the load and the energy storage unit. When the input power supply is powered down, the first switch of the primary-side switch unit maintains in an off state, the second switch of the secondary-side switch unit performs an on-off action, and the energy storage unit supplies power to the load.

Abstract: For a content item with unknown tasks performed by a viewing user on an online system, the online system predicts a likelihood of interacting with each content item using a prediction model associated with a plurality of tasks. The prediction model comprises a plurality of independent layers, a plurality of shared layers and a plurality of separate layers. Each independent layer is configured to extract features, for each task, that are not shared across the plurality of tasks. The plurality of shared layers are configured to extract common features that are shared across the plurality of tasks. Each separate layer is configured to predict likelihood of the viewing user performing a task associated with the separate layer based on the features extracted from the plurality of independent layers and the plurality of shared layers.