Abstract: A computing system is provided, including a plurality of processing devices configured to execute a Mixture-of-Experts (MoE) layer included in an MoE model. The MoE layer includes a plurality of expert sub-models that each have a respective plurality of parameter values. The MoE layer is configured to be switchable between a data parallel mode and an expert-data-model parallel mode without conveying the respective parameter values of the expert sub-models among the plurality of processing devices.

Abstract: A computing system including a plurality of processing devices configured to execute a Mixture-of-Experts (MoE) layer included in an MoE model. The processing devices are configured to execute the MoE layer at least in part by, during a first collective communication phase between the processing devices, splitting each of a plurality of first input tensors along a first dimension to obtain first output tensors. Executing the MoE layer further includes processing the first output tensors at a respective a plurality of expert sub-models to obtain a plurality of second input tensors. Executing the MoE layer further includes, during a second collective communication phase between the processing devices, receiving the second input tensors from the expert sub-models and concatenating the second input tensors along the first dimension to obtain second output tensors. Executing the MoE layer further includes outputting the second output tensors as output of the MoE layer.

Abstract: The present disclosure relates to a method and apparatus for knowledge graph construction, storage medium and electronic device. The method for knowledge graph construction, comprises: identifying an entity concept from a title text of a target web page and at least one entity corresponding to the entity concept from a body text of the target web page; constructing a syntax parse tree of the title text based on syntax parse rules of a language to which the title text belongs, and determining, from the syntax parse tree, a modifier for modifying the entity concept; and generating a knowledge graph based on the entity concept, the modifier, and the at least one entity. Through the solution of the present disclosure, knowledge graphs with high accuracy and high recall rates are constructed without structured processing on target web pages.

Abstract: A computing system including a plurality of processing devices configured to execute a Mixture-of-Experts (MoE) layer. The processing devices are configured to execute the MoE layer at least in part by receiving an input tensor including input tokens. Executing the MoE layer further includes computing a gating function output vector based on the input tensor and computing a sparse encoding of the input tensor and the gating function output vector. The sparse encoding indicates one or more destination expert sub-models. Executing the MoE layer further includes dispatching the input tensor for processing at the one or more destination expert sub-models, and further includes computing an expert output tensor. Executing the MoE layer further includes computing an MoE layer output at least in part by computing a sparse decoding of the expert output tensor. Executing the MoE layer further includes conveying the MoE layer output to an additional computing process.

Abstract: Aspects of the present disclosure are directed to systems and methods for detecting Partial Discharge (PD) associated with wide bandgap semiconductor-based electrical drive systems in real time. In one aspect, signals, including noise associated with drive switching and other background noises are detected using a sensing device. The signals are received by a real-time spectrum analyzer. The spectrum analyzer transforms the signals into the frequency domain and determines or registers the frequency domain profiles of the signals. The spectrum analyzer performs signal discrimination between at least one other signal included within the signals and switching noise based on their frequency domain profiles. Based on the discrimination analysis, the presence of a partial discharge signal may be detected. A physics-based signal discrimination approach can also be used for signal discrimination, for example utilizing pressure-dependency of characteristics of a PD signal.

Abstract: A power converter assembly includes a housing, a cold plate in the housing, and a set of power converter components on the cold plate. The cold plate is configured to receive a coolant. A first medium in the housing surrounds the cold plate and components. An expandable heat transfer structure is attached to one of the cold plate and the housing.

Abstract: An electrical filter circuit is disclosed. The circuit includes a first input line and a second input line. A first transmission line is coupled electrically in series at a first node with a first output line, and an inductor is coupled electrically in series between the first input line and the first transmission line. The filter also includes a second transmission line having a first impedance coupled to the first node. The second input line is coupled electrically in series at a second node with a second output line. A third transmission line is coupled to the second node, and a capacitor is coupled electrically in series between the second transmission line and the third transmission line. The first output line has a second impedance that is greater than the first impedance.

Abstract: Partial discharge detection techniques are provided. In one aspect, a method of detecting partial discharge in an electrical power system having a power electronics converter is provided. A first current signal is captured in response to a first applied voltage. A second current signal is captured in response to a second applied voltage, the second applied voltage being different than the first applied voltage. The first current signal is set as a reference signal. A difference signal is determined based on the second current signal and the reference signal. A determination is made as to whether partial discharge is present based on the difference signal. The method may iterate, and for each iteration, the reference signal is refreshed on a rolling basis as a previously captured current signal measured in response to a previously applied voltage. The applied voltage is also stepped up or down for each iteration.

Abstract: An electrical filter circuit is disclosed. The circuit includes a first input line and a second input line. A first transmission line is coupled electrically in series at a first node with a first output line, and an inductor is coupled electrically in series between the first input line and the first transmission line. The filter also includes a second transmission line having a first impedance coupled to the first node. The second input line is coupled electrically in series at a second node with a second output line. A third transmission line is coupled to the second node, and a capacitor is coupled electrically in series between the second transmission line and the third transmission line. The first output line has a second impedance that is greater than the first impedance.

Abstract: A cable that includes a conductor defining a hollow interior, a casing surrounding the conductor, an electrical insulator positioned between the conductor and the casing, and a fluid positioned within the hollow interior of the conductor.

Abstract: An RFID-based sensing method including receiving, from an RFID tag with a tuning circuit when the tuning circuit is arranged in a first state, a first RF signal; and receiving, from the RFID tag when the tuning circuit is arranged in a second state, a second RF signal. The tuning circuit is arranged to affect or facilitate impedance matching of an antenna and an integrated circuit of the RFID tag. The method also includes processing the first and second RF signals received to determine respective differences in two or more properties of the first and second RF signals received; and determining, based on the processing, a status associated with an environment in which the RFID tag is arranged.

Abstract: An RFID-based sensing method including receiving, from an RFID tag with a tuning circuit when the tuning circuit is arranged in a first state, a first RF signal; and receiving, from the RFID tag when the tuning circuit is arranged in a second state, a second RF signal. The tuning circuit is arranged to affect or facilitate impedance matching of an antenna and an integrated circuit of the RFID tag. The method also includes processing the first and second RF signals received to determine respective differences in two or more properties of the first and second RF signals received; and determining, based on the processing, a status associated with an environment in which the RFID tag is arranged.

Abstract: Aspects of the present disclosure are directed to systems and methods for detecting Partial Discharge (PD) associated with wide bandgap semiconductor-based electrical drive systems in real time. In one aspect, signals, including noise associated with drive switching and other background noises are detected using a sensing device. The signals are received by a real-time spectrum analyzer. The spectrum analyzer transforms the signals into the frequency domain and determines or registers the frequency domain profiles of the signals. The spectrum analyzer performs signal discrimination between at least one other signal included within the signals and switching noise based on their frequency domain profiles. Based on the discrimination analysis, the presence of a partial discharge signal may be detected. A physics-based signal discrimination approach can also be used for signal discrimination, for example utilizing pressure-dependency of characteristics of a PD signal.

Abstract: Embodiments are disclosed for a stereoscopic device (also referred to simply as the “device”) that captures three-dimensional (3D) images and videos with a wide field of view and provides a virtual reality (VR) experience by immersing a user in a simulated environment using the captured 3D images or videos.

Abstract: Embodiments are disclosed for embedding calibration metadata for a stereoscopic video capturing device. The device captures a sequence of stereoscopic images by a plurality of image sensors and combines the captured sequence of stereoscopic images into a stereoscopic video sequence. The device further embeds calibration information into the stereoscopic video sequence in a real time as the sequence of stereoscopic images is being recorded. The calibration information can be used to correct distortion caused by hardware variances of individual video capturing devices. The corrected stereoscopic videos can be used to provide a virtual reality (VR) experience by immersing a user in a simulated environment.

Abstract: Embodiments are disclosed for a stereoscopic device (also referred to simply as the “device”) that captures three-dimensional (3D) images and videos with a wide field of view and provides a virtual reality (VR) experience by immersing a user in a simulated environment using the captured 3D images or videos.

Abstract: Embodiments are disclosed for embedding calibration metadata for a stereoscopic video capturing device. The device captures a sequence of stereoscopic images by a plurality of image sensors and combines the captured sequence of stereoscopic images into a stereoscopic video sequence. The device further embeds calibration information into the stereoscopic video sequence in a real time as the sequence of stereoscopic images is being recorded. The calibration information can be used to correct distortion caused by hardware variances of individual video capturing devices. The corrected stereoscopic videos can be used to provide a virtual reality (VR) experience by immersing a user in a simulated environment.

Abstract: Embodiments are disclosed for a stereoscopic device (also referred to simply as the “device”) that captures three-dimensional (3D) images and videos with a wide field of view and provides a virtual reality (VR) experience by immersing a user in a simulated environment using the captured 3D images or videos.

Abstract: Embodiments are disclosed for a stereoscopic device (also referred to simply as the “device”) that captures three-dimensional (D) images and videos with a wide field of view and provides a virtual reality (VR) experience by immersing a user in a simulated environment using the captured 3D images or videos.