Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating data structures from graphs. The computer accesses a graph having patient nodes representing patients and patient health data nodes representing health data for patients, the nodes being connected by edges. The computer generates subgraphs by identifying patient nodes and patient health data nodes associated with a particular healthcare provider. The computer generates, by a subgraph neural network, a healthcare provider data structure for a respective subgraph. The computer generates, by a first graph neural network, patient data structures for a respective patient graph network and health data structures for a respective health data graph network. Each healthcare provider data structure, patient graph network, health data structure, has a lower dimension than the corresponding subgraph, patient graph network, and health data graph network, respectively.

Abstract: Provided are a silicon-based microphone device and an electronic device. The silicon-based microphone device comprises a circuit board, a shielding housing and at least two differential silicon-based microphone chips, wherein at least two sound inlet holes are provided on the circuit board, the shielding housing covers one side of the circuit board and forms a sound cavity with the circuit board, the silicon-based microphone chips are all located inside the sound cavity, the differential silicon-based microphone chips are respectively disposed at the sound inlet holes, and a back cavity of each differential silicon-based microphone chip is communicated with the sound inlet hole at the corresponding position, each of the differential silicon-based microphone chips comprises a first microphone structure and a second microphone structure, all of the first microphone structures are electrically connected, and all of the second microphone structures are electrically connected.

Abstract: Disclosed is a method for updating a node model that resists discrimination propagation in federated learning. The method includes: obtaining a node model corresponding to a data node; calculating a mean value of the distribution of class features and a quantity ratio corresponding to training data of the data node, calculating a distribution weighted aggregation model based on the node model, the mean value of the distribution of class features and the quantity ratio; calculating a regularization term corresponding to the data node based on the node model and the distribution weighted aggregation model; calculating a variance of the distribution of the class features corresponding to the data node, calculating a class balanced complementary term by using a cross-domain feature generator; and updating the node model based on the distribution weighted aggregation model, the regularization term, and the class balanced complementary term.

Abstract: Disclosed is a disentangled personalized federated learning method via consensus representation extraction and diversity propagation provided by embodiments of the present application. The method includes: receiving, by a current node, local consensus representation extraction models and unique representation extraction models corresponding to other nodes, respectively; extracting, by the current node, the representations of the data of the current node by using the unique representation extraction models of other nodes respectively, and calculating first mutual information between different sets of representation distributions, determining similarity of the data distributions between the nodes based on the size of the first mutual information, and determining aggregation weights corresponding to the other nodes based on the first mutual information; the current node obtains the global consensus representation aggregation model corresponding to the current node.

Abstract: Disclosed is a method for updating a node model that resists discrimination propagation in federated learning. The method includes: obtaining a node model corresponding to a data node; calculating a mean value of the distribution of class features and a quantity ratio corresponding to training data of the data node, calculating a distribution weighted aggregation model based on the node model, the mean value of the distribution of class features and the quantity ratio; calculating a regularization term corresponding to the data node based on the node model and the distribution weighted aggregation model; calculating a variance of the distribution of the class features corresponding to the data node, calculating a class balanced complementary term by using a cross-domain feature generator; and updating the node model based on the distribution weighted aggregation model, the regularization term, and the class balanced complementary term.

Abstract: Disclosed is a disentangled personalized federated learning method via consensus representation extraction and diversity propagation provided by embodiments of the present application. The method includes: receiving, by a current node, local consensus representation extraction models and unique representation extraction models corresponding to other nodes, respectively; extracting, by the current node, the representations of the data of the current node by using the unique representation extraction models of other nodes respectively, and calculating first mutual information between different sets of representation distributions, determining similarity of the data distributions between the nodes based on the size of the first mutual information, and determining aggregation weights corresponding to the other nodes based on the first mutual information; the current node obtains the global consensus representation aggregation model corresponding to the current node.

Abstract: A deep learning model implements continuous, lifelong machine learning (LML) based on a Bayesian neural network using an inventive framework including wide, deep, and prior components that employ diverse algorithms to leverage available real-world healthcare data differently to improve prediction performance. The outputs from each component of the framework are fed into a wide and shallow neural network and the posterior structure of the final model output may be utilized as a prior structure when the deep learning model is refreshed with new data in a deep learning process. Lifelong learning is implemented by dynamically integrating present learning from the wide and deep learning components with past learning from traditional tree models in the prior component into future predictions. Thus, the present Bayesian deep neural network-based LML model increases accuracy in identifying patient profiles by continuously learning, as new data become available, without forgetting prior knowledge.

Abstract: An automatic charging apparatus for a furnace, including: a hopper provided with an inlet and an outlet opening; a buffer bin in which a buffer channel is provided, one end of the buffer channel including a discharge port communicating with the outlet opening, the other end being in communication with a feeding inlet of the furnace; a pushing mechanism configured to push feedstock in the buffer channel sequentially into the feeding inlet of the furnace, the pushing mechanism including a drive and a pushing element, the drive driving the pushing element to perform a reciprocating movement in the buffer channel so the feedstock in the buffer channel is sequentially pushed into the feeding inlet via the reciprocating movement of the pushing element. The solution can implement automatic charging of feedstock in the furnace, reduce labor intensity of operators, and realize accurate control of charging amount of the feedstock.

Abstract: A metal melting furnace including a stirring device as includes: a furnace body defining a chamber for accommodating molten metal; the stirring device including a stirring disc and a drive device, the stirring disc including a disc body, a feedstock holding portion, and a feedstock inlet via which feedstock is replenished to the feedstock holding portion, a stirring rod being connected on the disc body, a plurality of vertically through openings being provided on the disc body and/or on the feedstock holding portion, the drive device being in drive connection to the stirring rod, the drive device being configured to drive the stirring rod to lift such that the stirring disc is immersed in or lifted out of the molten metal in the chamber, the feedstock holding portion being configured to hold the feedstock on the stirring disc such that the feedstock move together with the stirring disc.

Abstract: A micromachined capacitive flow sensor includes a movable membrane having one or more venting holes and a perforated backplate having perforation holes or through holes. A gas gap (such as an air gap) is formed between the movable membrane and the perforated backplate. The movable membrane, the gas gap and the perforated backplate form a variable capacitor whose capacitance varies with a movement of the membrane relative to the perforated backplate. The sensor may be used to manufacture a packaged flow sensor product which may find numerous applications, for example, using the product as a switch to turn on and off the electric power to the heating elements of the aerosol delivery device in response to the puff and/or smoking action of the user.

Abstract: A micromachined capacitive flow sensor includes a movable membrane having one or more venting holes and a perforated backplate having perforation holes or through holes. A gas gap (such as an air gap) is formed between the movable membrane and the perforated backplate. The movable membrane, the gas gap and the perforated backplate form a variable capacitor whose capacitance varies with a movement of the membrane relative to the perforated backplate. The sensor may be used to manufacture a packaged flow sensor product which may find numerous applications, for example, using the product as a switch to turn on and off the electric power to the heating elements of the aerosol delivery device in response to the puff and/or smoking action of the user.

Abstract: Embodiments of the present application provide a silicon-based microphone apparatus and an electronic device. The silicon-based microphone apparatus comprises: a circuit board provided with at least two sound inlet holes; a shielding cover covering one side of the circuit board to form an acoustic cavity; at least two silicon-based microphone chips both provided on one side of the circuit board and located in the acoustic cavity, back cavities of the silicon-based microphone chips being communicated with the sound inlet holes in one-to-one correspondence; and a differential control chip, microphone structures of all the silicon-based microphone chips being sequentially electrically connected and then being electrically connected to an input end of the differential control chip.

Abstract: Provided in the embodiments of the present application are a silicon-based microphone device and an electronic device. The silicon-based microphone device comprises: a circuit board, provided with at least two sound inlets; a shielding cover, covering one side of the circuit board to form an acoustic cavity; at least two differential silicon-based microphone chips, both being provided on one side of the circuit board and arranged within the acoustic cavity; rear cavities of the differential silicon-based microphone chips being in communication in a one-to-one correspondence with the sound inlets; and a separating element, arranged within the acoustic cavity, and separating the acoustic cavity into acoustic subcavities corresponding to the rear cavities of at least some of the adjacent differential silicon-based microphone chips.

Abstract: Embodiments of the present application provide a silicon based microphone apparatus and an electronic device. The silicon based microphone apparatus comprises: a circuit board provided with at least one sound inlet hole; a shielding cover covering one side of the circuit board to form an acoustic cavity; at least two differential silicon based microphone chips arranged on one side of the circuit board and located within the acoustic cavity, back cavities of some differential silicon based microphone chips being communicated with the sound inlet holes in one-to-one correspondence; and a separator located within the acoustic cavity and used for separating the acoustic cavity into sub-acoustic cavities corresponding to the back cavities of at least some adjacent differential silicon based microphone chips.

Abstract: Provided are a silicon-based microphone device and an electronic apparatus. The silicon-based microphone device comprises: a circuit board, provided with at least two sound inlets; a shielding housing, covering one side of the circuit board; an even number of differential silicon-based microphone chips, which are arranged within an acoustic cavity, between every two differential silicon-based microphone chips, a first microphone structure of either differential silicon-based microphone chip being electrically connected to a second microphone structure of the other differential silicon-based microphone chip, and the second microphone structure of either differential silicon-based microphone chip being electrically connected to the first microphone structure of the other differential silicon-based microphone chip; and, a mounting plate, provided with at least one opening, the opening being in communication with some of the sound inlets.

Abstract: An audio capturing device comprises: a housing and a silicon-based microphone device arranged within the housing. The silicon-based microphone device comprises a circuit board and an even number of silicon-based microphone chips provided at one side of the circuit board. The circuit board is provided with at least one sound inlet. The at least one sound inlet is in communication in a one-to-one correspondence to rear cavities of some of the silicon-based microphone chips of the even number of silicon-based microphone chips. Sound channels in communication in a one-to-one correspondence to sound inlets are provided in the housing. The corresponding rear cavities, sound inlets, and sound channels form first acoustic cavities. The rear cavities form second acoustic cavities. The first acoustic cavities are different from the second acoustic cavities in terms of volume and/or shape.

Abstract: A sound collection device comprises a housing and a silicon-based microphone device located in the housing. The silicon-based microphone device comprises a circuit board and an even number of silicon-based microphone chips provided on one side of the circuit board; the circuit board is provided with an even number of sound inlet holes, and the even number of sound inlet holes have one-to-one correspondence to back cavities of the even number of silicon-based microphone chips; the housing is provided with sound channels in communication with the sound inlet holes in one-to-one correspondence; the correspondingly communicated back cavities and the sound inlet holes form acoustic cavities; or, the correspondingly communicated back cavities, sound inlet holes, and sound channels form acoustic cavities; and at least two acoustic cavities have different volumes and/or shapes.

Abstract: Provided are a silicon-based microphone device and an electronic apparatus. The silicon-based microphone device comprises: a circuit board, wherein at least two sound inlets are formed on the circuit board; a shielding housing that covers one side of the circuit board; an even number of differential silicon-based microphone chips that all are located in a sound cavity, wherein in each two differential silicon-based microphone chips, the first microphone structure of one differential silicon-based microphone chip is electrically connected to the second microphone structure of the other differential silicon-based microphone chip, and the second microphone structure of said one differential silicon-based microphone chip is electrically connected to the first microphone structure of said other differential silicon-based microphone chip; and a mounting plate, wherein an even number of holes communicated with the sound inlets are formed on the mounting plate.

Abstract: Provided are a silicon-based microphone device and an electronic device. The silicon-based microphone device comprises a circuit board, a shielding housing and at least two differential silicon-based microphone chips, wherein at least two sound inlet holes are provided on the circuit board, the shielding housing covers one side of the circuit board and forms a sound cavity with the circuit board, the silicon-based microphone chips are all located inside the sound cavity, the differential silicon-based microphone chips are respectively disposed at the sound inlet holes, and a back cavity of each differential silicon-based microphone chip is communicated with the sound inlet hole at the corresponding position, each of the differential silicon-based microphone chips comprises a first microphone structure and a second microphone structure, all of the first microphone structures are electrically connected, and all of the second microphone structures are electrically connected.