Abstract: Methods, systems, and circuits for training a neural network include applying noise to a set of training data across wordlines using a respective noise switch on each wordline. A neural network is trained using the noise-applied training data to generate a classifier that is robust against adversarial training.

Abstract: Disclosed herein are compounds of formula I and/or a stereoisomer, stable isotopologue, and/or pharmaceutically acceptable salts thereof; and therapeutic uses of these compounds, which are inhibitors of tryptophan 2, 3-dioxygenase 2 (TDO2) and/or indoleamine 2, 3-dioxygenase 1 (IDO1), potentially useful in the treatment of diseases treatable, such as cancers.

Abstract: Techniques facilitating matrix multiplication on a systolic array are provided. A computer-implemented method can comprise populating, by a system operatively coupled to a processor, respective first registers of one or more processing elements of a systolic array structure with respective input data bits of a first data matrix. The one or more processing elements can comprise a first processing element that comprises a first input data bit of the first data matrix and a first activation bit of a second data matrix. The method can also include determining, by the system, at the first processing element, a first partial sum of a third data matrix. Further, the method can include streaming, by the system, the first partial sum of the third data matrix from the first processing element.

Abstract: (I) Disclosed herein are compounds of formula I, and/or a stereoisomer, a stable isotope, or a pharmaceutically acceptable salt thereof; and therapeutic uses of these compounds, which are kinase inhibitors potentially useful in the treatment of diseases treatable, such as cancers, as well as pharmaceutical compositions, comprising a compound of formula I, and/or a stereoisomer, a stable isotope, or a pharmaceutically acceptable salt thereof, and a pharmaceutically acceptable carrier.

Abstract: A tadalafil orally dissolving film and a preparation method thereof are provided. The tadalafil orally dissolving film consisting of specific components is prepared through a conventional process. Particularly, the tadalafil orally dissolving film is prepared by tadalafil, film-forming material, plasticizer, correctant, colorant and aromatic, wherein the film-forming material includes components of: (a) at least one member selected from polysaccharide; (b) at least one member selected from organic acid; and (c) at least one member selected from hydroxyalkyl cellulose. Compared with prior art, the tadalafil orally dissolving film provided by the present invention has an easy preparation process, no specific requirement on production devices and a storing stability, and is easy to reproduce, has a low requirement on storage conditions, and is easy to package, transport and store.

Abstract: Disclosed herein is a compound of formula (I) and/or a pharmaceutically acceptable salt thereof that can serve as Erk inhibitors. They are potentially useful in the treatment of diseases treatable by inhibition of Erk, such as cancers. Also disclosed herein is a pharmaceutical composition, comprising a compound of formula I and/or a pharmaceutically acceptable salt thereof and a pharmaceutically acceptable carrier.

Abstract: A power transmission system for a vehicle and a vehicle including the same are provided. The power transmission system includes an engine unit configured to generate power, a transmission unit adapted to selectively be coupled with the engine unit, and configured to transmit the power generated by the engine unit, a first motor generator coupled with the transmission unit, an output unit configured to transmit the power output by the transmission unit to at least one of front and rear wheels of the vehicle, and a second motor generator configured to drive the at least one of the front and rear wheels.

Abstract: A mold with cavities for nanoimprint lithography and a method of imprinting the same. The mold is a one-piece structure, including an upper mold and a lower mold which are welded together. The upper mold is evenly provided with a plurality of stepped holes. A plurality of grooves are provided at an upper surface of the lower mold. Both ends of each groove are respectively connected to an external high-pressure water pump and an external vacuum pump via pipelines. An airtight cavity is additionally provided at the interior of the mold to form a negative pressure in the cavity by vacuumizing during the imprint process. In the demolding, a crack is formed at the interface between the mold and the imprint resist by using water pressure and expands to the entire interface for the demolding of the imprinted structure.

Abstract: A method for improving performance of a predefined Deep Neural Network (DNN) convolution processing on a computing device includes inputting parameters, as input data into a processor on a computer that formalizes a design space exploration of a convolution mapping, on a predefined computer architecture that will execute the predefined convolution processing. The parameters are predefined as guided by a specification for the predefined convolution processing to be implemented by the convolution mapping and by a microarchitectural specification for the processor that will execute the predefined convolution processing. The processor calculates performance metrics for executing the predefined convolution processing on the computing device, as functions of the predefined parameters, as proxy estimates of performance of different possible design choices to implement the predefined convolution processing.

Abstract: Techniques facilitating matrix multiplication on a systolic array are provided. A computer-implemented method can comprise populating, by a system operatively coupled to a processor, respective first registers of one or more processing elements of a systolic array structure with respective input data bits of a first data matrix. The one or more processing elements can comprise a first processing element that comprises a first input data bit of the first data matrix and a first activation bit of a second data matrix. The method can also include determining, by the system, at the first processing element, a first partial sum of a third data matrix. Further, the method can include streaming, by the system, the first partial sum of the third data matrix from the first processing element.

Abstract: Techniques facilitating matrix multiplication on a systolic array are provided. A computer-implemented method can comprise populating, by a system operatively coupled to a processor, respective first registers of one or more processing elements of a systolic array structure with respective input data bits of a first data matrix. The one or more processing elements can comprise a first processing element that comprises a first input data bit of the first data matrix and a first activation bit of a second data matrix. The method can also include determining, by the system, at the first processing element, a first partial sum of a third data matrix. Further, the method can include streaming, by the system, the first partial sum of the third data matrix from the first processing element.

Abstract: Disclosed herein is a compound of formula I and/or a pharmaceutically acceptable salt thereof that can serve as Erk inhibitors: They are potentially useful in the treatment of diseases treatable by inhibition of Erk, such as cancers. Also disclosed herein is a pharmaceutical composition comprising a compound of formula I and/or a pharmaceutically acceptable salt thereof and a pharmaceutically acceptable carrier.

Abstract: The present invention discloses a locking device, a power assembly, a power transmission system and a vehicle. The locking device includes: a first flange, the first flange being adapted to be fixed on the first shaft; a second flange, the second flange being adapted to be fixed on the second shaft; a synchronizing ring, the synchronizing ring being normally connected to the first flange to be adapted to rotate synchronously with the first flange, and the synchronizing ring being slidable relative to the first flange; and a driving component, the driving component selectively pushing the synchronizing ring to slide from an unlocked position to a locked position in an axial direction of the first flange. When the synchronizing ring is in the locked position, the synchronizing ring is connected to the second flange be adapted to rotate the second flange synchronously with the first flange. When the synchronizing ring is in the unlocked position, the synchronizing ring is separated from the second flange.

Abstract: A locking apparatus, a powertrain, a power transmission system, and a vehicle, the locking system (200) comprising: a first flange (201a) and a second flange (201b); and first and second flange locking structures (Q1, Q2), the first and second flange locking structures (Q1, Q2) both being used for selectively locking the first flange (201a) and the second flange (201b) in order to make the second flange (201b) rotate synchronously with the first flange (201a), or in order to make the first flange (201a) rotate synchronously with the second flange (201b); the first and second flange locking structures (Q1, Q2) both comprise: a synchronising ring (202), the synchronising ring (202) being constantly connected to the corresponding flange in order to rotate synchronously with the corresponding flange, and the synchronising ring (202) being able to slide relative to the corresponding flange; and a drive assembly (203), the drive assembly (203) being capable of selectively pushing the synchronising ring to slide alo

Abstract: Techniques facilitating matrix multiplication on a systolic array are provided. A computer-implemented method can comprise populating, by a system operatively coupled to a processor, respective first registers of one or more processing elements of a systolic array structure with respective input data bits of a first data matrix. The one or more processing elements can comprise a first processing element that comprises a first input data bit of the first data matrix and a first activation bit of a second data matrix. The method can also include determining, by the system, at the first processing element, a first partial sum of a third data matrix. Further, the method can include streaming, by the system, the first partial sum of the third data matrix from the first processing element.

Abstract: A power transmission system for a vehicle and a vehicle including the same are provided.

Abstract: A mold with cavities for nanoimprint lithography and a method of imprinting the same. The mold is a one-piece structure, including an upper mold and a lower mold which are welded together. The upper mold is evenly provided with a plurality of stepped holes. A plurality of grooves are provided at an upper surface of the lower mold. Both ends of each groove are respectively connected to an external high-pressure water pump and an external vacuum pump via pipelines. An airtight cavity is additionally provided at the interior of the mold to form a negative pressure in the cavity by vacuumizing during the imprint process. In the demolding, a crack is formed at the interface between the mold and the imprint resist by using water pressure and expands to the entire interface for the demolding of the imprinted structure.

Abstract: The present disclosure discloses an electric vehicle, an active safety control system of an electric vehicle, and a control method of the active safety control system of an electric vehicle. The electric vehicle includes: multiple wheels, multiple motors, a wheel speed detection module, a steering wheel rotation angle sensor, a yaw rate sensor, and a battery pack. The active safety control system includes: an acquisition module, acquiring the wheel speed signal, the direction information, the yaw information, status information of the battery pack, and status information of the multiple motors; a status determining module, determining status of the electric vehicle; and a control module, generating a control instruction and delivering the control instruction to at least one motor.

Abstract: A power transmission system of a vehicle and a gear-shift control method are provided. The power transmission system includes a first drive assembly (101) and a second drive assembly (102). The first drive assembly includes a first automatic transmission (2) and a first motor (1) connected with the first automatic transmission (2) for outputting power to two wheels of the vehicle. The second drive assembly (102) includes a second automatic transmission (2000) and a second motor (1000) connected with the second automatic transmission (2000) for outputting power to the other two wheels of the vehicle. The first and second automatic transmission are configured to be shifted to a preset gear or a gear adjacent to the preset gear, such that when one of the first and second automatic transmission is shifted to the preset gear, the other one of the first and second automatic transmission is shifted to the preset gear or the gear adjacent to the preset gear.

Abstract: Techniques facilitating matrix multiplication on a systolic array are provided. A computer-implemented method can comprise populating, by a system operatively coupled to a processor, respective first registers of one or more processing elements of a systolic array structure with respective input data bits of a first data matrix. The one or more processing elements can comprise a first processing element that comprises a first input data bit of the first data matrix and a first activation bit of a second data matrix. The method can also include determining, by the system, at the first processing element, a first partial sum of a third data matrix. Further, the method can include streaming, by the system, the first partial sum of the third data matrix from the first processing element.