Abstract: Disclosed in the present invention is a differentiable method for mining constant rules. On the basis of defining constant operators and path operators according to a knowledge graph, a fusion attention mechanism is used to evaluate relations passed by rules by using attention values. Meanwhile, attentions are calculated by aggregating surrounding attributes and the corresponding attribute values for tail node of each hop respectively. Attentions of the attributes are used to enhance selection of relations in the rules to achieve link prediction. According to generated model parameters, high quality symbolic rules are output by parameter analysis and statistics. This method is particularly suitable for application scenarios with complex reasoning requirements that require high prediction accuracy, provide interpretations for predictions, and need to precipitate reasoning rules.

Abstract: Provided is a solar panel. The solar panel includes a frame having an accommodation opening, a front plate located at one of two opposite sides of the frame, a back plate located at another one of the two opposite sides of the frame, and a cell. The accommodation opening is enclosed by the front plate, the back plate, and the frame to form an enclosed space, and the cell is disposed in the enclosed space.

Abstract: The present invention relates to a carbide tool cleaning and coating production line and a method, including a cleaning device including a support frame, a cleaning mechanism and a drying mechanism are sequentially disposed under the support frame connected to a moving mechanism, the moving mechanism is connected to a lifting mechanism being capable of being connected to a tool fixture bracket being configured to accommodate the tool fixture; a coating device including a coating chamber which a plane target mechanism and a turntable assembly disposed in, the turntable assembly is capable of being connected to a plurality of tool fixtures being capable of rotating around an axial line of the coating chamber under the driving of the turntable assembly and rotating around an axial line thereof at the same time; and, a manipulator being disposed between the cleaning device and the coating device.

Abstract: The present invention discloses a protein transformation method based on an amino acid knowledge graph and active learning, including: building an amino acid knowledge graph based on biochemical attributes of amino acids; enhancing protein data in combination with the amino acid knowledge graph to obtain enhanced protein data, and performing representation learning to obtain first enhanced protein representations; performing representation learning on the protein data or the protein data and the amino acid knowledge graph by using a pre-trained protein model to obtain second enhanced protein representations; synthesizing the first enhanced protein representations and the second enhanced protein representations to obtain enhanced protein representations; taking the enhanced protein representations as samples, and through active learning, screening out representative samples from the samples, manually annotating protein properties, and training a protein property prediction model by using the manually annotated

Abstract: Disclosed in the present invention is a generative event extraction method based on ontology guidance, including: (1) constructing an event ontology knowledge base; (2) designing an event trigger word extraction template and an event argument extraction template; mapping an input event text to a first input sequence, and mapping an input event text integrating an event ontology to a second input sequence; (3) designing a class label mapping function that maps multi-word labels to event types and/or role types; (4) extracting the event ontology corresponding to the input event from the event ontology knowledge base, and constructing the first input sequence and the second input sequence according to the event trigger word extraction template and the event argument extraction template; and (5) predicting, by the event extraction model, the event type and the role type according to the class label mapping function and a processing mechanism thereof, and outputting an event trigger word span and an event argument

Abstract: The present invention discloses a perceptual representation learning method for protein conformations based on a pre-trained language model, including: obtaining a protein made up of an amino acid sequence, building different data sets according to protein conformations, and defining a prompt for each type of protein conformation; building, based on a pre-trained language model, a representation learning module for fusing an embedding representation of each type of the prompt into an embedding representation of the protein, so as to obtain a protein embedding representation under a prompt identifier; building a task module for performing task prediction on a task corresponding to each type of protein conformation based on the protein embedding representation under the prompt identifier; building a loss function for each type of task based on a task prediction result and a tag, and updating model parameters of the representation learning module and the task module in combination with loss functions of all type

Abstract: A laterally diffused metal oxide semiconductor device and a preparation method thereof are disclosed. The semiconductor device includes: a substrate; a body region having a first conductivity type and formed in the substrate; a drift region, having a second conductivity type, formed in the substrate and adjacent to the body region; a field plate structure, formed on the drift region, a lower surface of an end of the field plate structure close to the body region being flush with the upper surface of the substrate, and the end of the field plate structure close to the body region also having an upwardly extending inclined surface; and a drain region, having a second conductivity type, formed in an upper layer of the drift region, and in contact with the end of the field plate structure away from the body region.

Abstract: A multi-procedure integrated automatic production line for hard alloy blades under robot control is provided. The production line includes a rail-guided robot. A cutter passivation device and a blade cleaning and drying device are arranged on one side of the rail-guided robot. A blade-coating transfer table, a blade coating device, a blade boxing transfer table, a blade-tooling dismounting device and a blade boxing device are sequentially arranged on another side of the rail-guided robot. The blade-tooling dismounting device is arranged on one side of the blade boxing transfer table. The production line further includes squirrel-cage toolings for carrying the blades. The squirrel-cage tooling that are loaded with the blades can run among the cutter passivation device, the blade cleaning and drying device, the blade-coating transfer table and the blade boxing transfer table. The blades after being treated through the blade-tooling dismounting device are sent to the blade boxing device.

Abstract: A preparation method of a zeolite/polyimide composite membrane includes: synthesizing a zeolite-doped polyamic acid precursor casting solution by condensation polymerization synthesis; coating a substrate with the obtained casting solution, and obtaining a zeolite/polyamic acid composite porous membrane by non-solvent induced phase separation; and obtaining the zeolite/polyimide composite membrane by performing thermal imidization on the zeolite/polyamic acid composite porous membrane through gradient heating.

Abstract: The present invention discloses an adaptive knowledge graph representation learning method for integrating a graph structure with text information, including: (1) sampling a neighbor triple of each of a head entity and a tail entity in a target triple; (2) calculating semantic representations of the target triple, and neighbor triples of its head and tail entities; (3) calculating structure representations of the head and tail entities of the target triple; (4) splicing the semantic representation of the target triple with the structure representations of its head and tail entities, inputting a spliced result into an adaptive classification layer, and calculating a classification result and a classification loss; and (5) optimizing the foregoing module based on an optimization algorithm of gradient descent, until a loss value converges, to obtain a final spliced result between the semantic representation of the target triple and the structure representations of its head and tail entities.

Abstract: The present invention is a molecular graph representation learning method based on contrastive learning, the method comprising: obtaining a molecular fingerprint representation of each molecule, and calculating a similarity between each two molecular fingerprints; collecting a full amount of chemical functional group information, and matching a corresponding functional group for each atom in the molecule; using a heterogeneous graph to model a molecular graph; using a RGCN in the structure-aware molecular encoder to encode the representation of each atom in the molecule and the representation of the functional group to which the atom belongs, and mapping the molecule to a feature space through an aggregation function to obtain a structure-aware feature representation; according to the fingerprint similarity between molecules, selecting positive and negative samples, and carrying out a comparative learning in the feature space; obtaining the structure-aware molecular encoder by using the contrastive learning m

Abstract: The present invention is a combined commodity mining method based on knowledge graph rule embedding, comprising: expressing rules, commodities, attributes, and attribute values as embeddings; splicing and inputting the embeddings of the rules and the embeddings of the attributes into a first neural network to obtain a importance scores of the attributes; splicing and inputting the rules and attributes into a second neural network to obtain the embeddings of the attribute values that the rules should take under the attributes; calculating a similarity between the value of two inputted commodities under the attribute and the embedding of the attribute value calculated by a model; after calculating scores of all attribute-attribute value pairs, summing up to obtain scores of these two commodities under this rule; then making the cross entropy loss with the real scores of these two commodities, and iteratively training based on an optimization algorithm having gradient descent; after the model is trained, parsing

Abstract: Disclosed is a four-way reversing valve, including a valve body, a motor transmission structure, and a reversing valve disc; the valve body includes a valve seat, an insulation sleeve welded to upper end surface of the valve seat, a D pipe, an S pipe, a C pipe, and an E pipe connected to lower end surface of the valve seat; the motor transmission structure includes motor worm gear and shaft transmission assembly and a planetary gearbox kit; the motor worm gear and shaft transmission assembly is sleeved onto upper external wall of the insulation sleeve through shaft hole of an external drive sprocket, and a non-through blind-hole is formed in a valve seat center; a central shaft is embedded into the blind-hole, and the central shaft connects the reversing valve disc and the planetary gearbox kit; the reversing valve disc and the planetary gearbox kit are inside the insulation sleeve.

Abstract: Disclosed is a four-way reversing valve, including a valve body, a motor transmission structure, and a reversing valve disc; the valve body includes a valve seat, an insulation sleeve welded to upper end surface of the valve seat, a D pipe, an S pipe, a C pipe, and an E pipe connected to lower end surface of the valve seat; the motor transmission structure includes motor worm gear and shaft transmission assembly and a planetary gearbox kit; the motor worm gear and shaft transmission assembly is sleeved onto upper external wall of the insulation sleeve through shaft hole of an external drive sprocket, and a non-through blind-hole is formed in a valve seat center; a central shaft is embedded into the blind-hole, and the central shaft connects the reversing valve disc and the planetary gearbox kit; the reversing valve disc and the planetary gearbox kit are inside the insulation sleeve.

Abstract: The present disclosure relates to a SDN-based method for mirroring packets, wherein a SDN controller is coupled to an upper layer application and at least one data switching exchange respectively, and the method including: a) the upper layer application sends a mirroring instruction to the SDN controller through a first northbound interface of the SDN controller; b) the SDN controller generates a second flow table based on the mirroring instruction and a first flow table sent by a first data switching exchange; wherein the first data switching exchange initiates transmission of the packets, the first flow table encapsulates the packets, and the second flow table includes at least an action command corresponding to the mirroring instruction; and c) a second data switching exchange extracts the packets from the second flow table, and mirrors the packets to the designated node based on the action command.

Abstract: An electronic expansion valve with high flow control precision, comprises a valve seat (1) and a valve rod (3). A through groove (11) penetrating the top and bottom of the valve seat is formed on the valve seat (1). A bearing seat (2) matching the through groove is inserted in the through groove (11). The bearing seat (2) includes a seat body (23), a through hole, (21) passing through the seat body, is formed in the seat body (3), and the through hole (21) matches the valve rod (3). An internal thread segment (24) is formed in the middle of the inner wall of the through hole (21). A matched straight tube is welded to the bottom of the through groove (11). The outer wall of the valve seat (1) is provided with a side hole communicated with the through groove (11), and a matching elbow pipe (4) is welded to the side hole. An isolating sleeve (10) is welded to the outer wall of the valve seat (1).

Abstract: An electronic expansion valve with high flow control precision, comprises a valve seat (1) and a valve rod (3). A through groove (11) penetrating the top and bottom of the valve seat is formed on the valve seat (1). A bearing seat (2) matching the through groove is inserted in the through groove (11). The bearing seat (2) includes a seat body (23), a through hole, (21) passing through the seat body, is formed in the seat body (3), and the through hole (21) matches the valve rod (3). An internal thread segment (24) is formed in the middle of the inner wall of the through hole (21). A matched straight tube is welded to the bottom of the through groove (11). The outer wall of the valve seat (1) is provided with a side hole communicated with the through groove (11), and a matching elbow pipe (4) is welded to the side hole. An isolating sleeve (10) is welded to the outer wall of the valve seat (1).

Abstract: The present disclosure relates to a SDN-based method for mirroring packets, wherein a SDN controller is coupled to an upper layer application and at least one data switching exchange respectively, and the method including: a) the upper layer application sends a mirroring instruction to the SDN controller through a first northbound interface of the SDN controller; b) the SDN controller generates a second flow table based on the mirroring instruction and a first flow table sent by a first data switching exchange; wherein the first data switching exchange initiates transmission of the packets, the first flow table encapsulates the packets, and the second flow table includes at least an action command corresponding to the mirroring instruction; and c) a second data switching exchange extracts the packets from the second flow table, and mirrors the packets to the designated node based on the action command.

Abstract: An artificial lift system for use with a subterranean well can include a cylinder having a piston reciprocably disposed therein, the piston having opposing sides, one side being selectively communicable with a hydraulic pressure source and a hydraulic reservoir, and the other side being selectively communicable with the hydraulic pressure source and at least one accumulator, and the accumulator being selectively communicable with an input side of a pump of the hydraulic pressure source. A method of controlling an artificial lift system can include connecting a cylinder to a hydraulic pressure source including a hydraulic pump, the pump being connected between the cylinder and at least one accumulator, the accumulator being connected to a gas pressure source, and operating a gas compressor of the gas pressure source, thereby increasing hydraulic pressure applied to the pump from the accumulator.

Abstract: An artificial lift system for use with a subterranean well can include a cylinder having a piston reciprocably disposed therein, the piston having opposing sides, one side being selectively communicable with at least one accumulator, and the other side having two areas, each of the areas being selectively communicable with a hydraulic pressure source and a hydraulic reservoir, and a gas pressure source connected to the accumulator, the gas pressure source including a gas compressor connected between at least one gas container and the accumulator. A method of controlling an artificial lift system can include connecting a cylinder to a hydraulic pressure source and to at least one accumulator, the accumulator being connected to a gas pressure source, and operating a gas compressor of the gas pressure source, thereby increasing hydraulic pressure applied to the cylinder from the accumulator.