30results about How to "Effective alignment" patented technology

The invention provides a robust multi-modal image segmentation method and system based on instance perception query, and the method comprises the steps: obtaining original image data and text description data, and generating training data through semantic analysis and causal graph construction; extracting image features and text features by adopting a multi-layer coding network, and establishing a feature dependency relationship through causal reasoning; generating instance clustering features based on attention calculation, and obtaining fusion features in combination with a multi-scale fusion strategy; and calculating query confidence and generating candidate segmentation masks, and outputting a final segmentation result through post-processing optimization. According to the method, by introducing causal reasoning and instance awareness query mechanisms, the cross-modal feature alignment precision and the robustness of the segmentation result are improved, and the method has better segmentation performance and generalization ability in a complex scene.

The invention discloses a method and device for adjusting the posture of a vehicle in a battery swap station, the battery swap station and the vehicle, and belongs to the technical field of battery charging and swap. The method comprises the following steps: controlling a passing platform of the battery swap station to rise to a first preset height, so that a plurality of probes arranged on the passing platform are in contact with a plurality of probe detection surfaces at the bottom of a vehicle; and adjusting the posture of the vehicle according to the measured value measured by each probe and a pre-calibrated calibration value. According to the embodiment of the invention, the plurality of probes are matched with the plurality of probe detection surfaces at the bottom of the vehicle, so that the measurement data of the plurality of contact points between the probes and the probe detection surfaces are sensed, the levelness deviation of the vehicle can be accurately identified by comparing the measurement values measured by the probes with the pre-calibrated calibration values, and the attitude of the vehicle is adjusted according to the levelness deviation. Therefore, the vehicle attitude levelness is improved, and the battery replacement success rate is improved.

The invention discloses a feeding mechanism of an automatic cold header for machining piston pins. The feeding mechanism comprises a main machine (1), a feeding mechanism body (20) is arranged on the front portion of the main machine (1), the feeding mechanism is connected with an iron wire frame (2) through an iron wire, and a discharging mechanism (3) is arranged on one side of the main machine (1). The feeding mechanism body comprises a base (4), fixed wheels (5) are arranged on one side of the base (4), and moving wheels (6) are arranged on the other side of the base (4). Side bases (7) are arranged on two sides of the base (4), sliding bases (8) are arranged on the inner sides of the side bases (7), and the contact surfaces of the sliding bases (8) and the side bases (7) are in sliding fit. The moving wheels (6) are arranged on the upper sides of the sliding bases (8), a threaded barrel (9) is fixed below the sliding bases (8), a threaded column (10) is arranged in the threaded barrel (9), a bearing pedestal (11) is arranged at one end of the threaded column (10), a rotating ring (12) is fixed to the other end of the threaded column (10), the rotating ring (12) is provided with a circle of small holes (13), and a rotating handle (14) is arranged in one small hole (13) in an embedded mode.

The invention provides a data alignment method and system and a storage medium. The method comprises the steps that world time information is acquired, and a constraint mode of a trigger signal is selected; outputting a trigger signal to a corresponding sensing device according to the world time information and the constraint mode, so that each sensing device obtains sensing data based on the trigger signal; and based on the constraint mode, selecting a corresponding alignment mode to perform alignment processing on the sensing data. According to the method provided by the invention, the problem that a back-end perception fusion application cannot obtain accurate perception information due to the fact that the perception contents of the perception devices are not matched is solved, and the perception data collected by the multiple perception devices are effectively aligned.

The invention discloses a multi-source domain rolling bearing fault diagnosis method and system based on a diffusion model. The diagnosis method comprises the steps that rolling bearing vibration signals under different rotating speed working conditions are collected; in the forward process of the diffusion model, converting a source domain label into a one-dimensional vector, and adding noise to destroy original distribution to obtain yT approximate to Gaussian distribution; replacing yT with a Gaussian distribution variable # imgabs0 # in a reverse process, learning noise distribution characteristics added in forward distribution through a noise prediction network, denoising # imgabs1 # step by step to obtain # imgabs2 # in the same distribution as y0, inputting source domain data into the noise prediction network to extract data characteristics, inputting target domain data into the noise prediction network, and outputting the target domain data into the noise prediction network; global noise priori and local noise priori corresponding to the target domain data are obtained; and inputting source domain data and target domain data into a noise prediction network, mapping source domain data features to a target domain by using a multi-source domain adaptive algorithm, learning noise distribution features related to the target domain data, and gradually denoising # imgabs3 # to obtain a fault type of the target domain data.

The present application relates to the field of sheet media processing, and particularly relates to a deviation rectifying device, a sheet media processing device and a control method of the deviation rectifying device; a rack of the deviation rectifying device is provided with a conveying channel, the conveying channel is used for conveying sheet media, and one side of the conveying channel is provided with a reference surface; a support frame of the deviation rectifying device is rotationally connected with the rack and has a first position and a second position, and a deviation rectifying wheel is rotationally arranged on the support frame and extends into the conveying channel; a first driving mechanism is used for driving the support frame to rotate between the first position and the second position; a second driving mechanism is used for driving the deviation rectifying wheel to rotate; when the support frame is located at the first position, an axis of the deviation rectifying wheel is arranged at a first included angle with the reference surface; when the support frame is located at the second position, an axis of the deviation rectifying wheel is arranged at a second included angle with the reference surface, and the first included angle and the second included angle are both less than or equal to 90°. The deviation rectifying device can improve the success rate of deviation rectification of sheet media.

The invention provides a magnetic clamping auxiliary laser centering device for a drilling machine. Comprising a laser transmitter, a magnetic shell, an end cover, a bolt, a sealing cover, a supporting block, a clamping claw and a limiting ring. The supporting block is of a split type combined structure and comprises a first supporting block body, a second supporting block body and a third supporting block body, the lower portion of the third supporting block body extends out of the magnetic shell, a laser transmitter is assembled in the center of the third supporting block body, and the laser transmitter and the center of the supporting block are collinear; the gripper jaw is arranged in a spacing cavity of the supporting block, and repulsive magnetic force is in contact with the drill bit and centered and locked; the limiting ring is assembled on the supporting block to limit the clamping jaw; the end cover is concentric with the magnetic shell and assembled at one end of the magnetic shell, the supporting blocks are contained in the magnetic shell, and the laser transmitter assembled at the extending position of the third supporting block extends out of the magnetic shell. The sealing cover is concentric with the magnetic shell and assembled at the other end of the magnetic shell; and the bolt is used for detachably assembling the end cover and the sealing cover on the magnetic shell. The center point and the center line of the drill bit are indicated through laser.

The invention discloses a power transmission line infrared image instance segmentation method and device, electronic equipment and a storage medium, and the method comprises the steps: carrying out the training of a first teacher model according to a visible light image and label information, and generating a first trained teacher model; according to the first infrared image and the label information, training a second teacher model to generate a second trained teacher model; according to the second infrared image and the first pseudo label information, training the first student model to generate a first trained student model; according to the second infrared image and the second pseudo label information, training the second student model to generate a second trained student model; according to the second infrared image and the third pseudo label information, performing fine tuning training on the first trained student model and the second trained student model to generate a second convergence teacher model; and performing instance segmentation on the infrared image of the power transmission line through the second convergence teacher model. According to the invention, the instance segmentation precision of the infrared image can be improved.

The invention belongs to the technical field of bearing lubrication state monitoring, and particularly relates to a bearing lubrication state real-time monitoring method based on hydrodynamic model correction, and the method comprises the steps: obtaining an original vibration acceleration signal and operation state data, carrying out the variational mode decomposition of the original vibration acceleration signal, and extracting a cavitation feature component; calculating a cavitation kinetic energy impact index according to the extracted cavitation characteristic component and the operation state data; dynamically correcting the equivalent fluid density of the lubricating oil according to the cavitation kinetic energy impact index and the contact area pressure of the bearing estimated by the running state data; and solving the minimum oil film thickness of a lubricating oil film by utilizing the corrected equivalent fluid density, and determining a lubrication safety margin index by combining the temperature rise rate of the bearing. According to the invention, the defect of monitoring result distortion caused by single-phase fluid hypothesis is overcome, and the accuracy of bearing lubrication state monitoring and early risk early warning is improved.

The invention discloses an aircraft rapid alignment system, method and equipment based on a starlight navigation device and a medium, and relates to the field of aircraft alignment, the system comprises an inertial navigation system, a GNSS satellite navigation receiving module, an angle rotating mechanism, the starlight navigation device, a light source and a processing module; the inertial navigation system, the GNSS satellite navigation receiving module and the angle rotating mechanism are fixedly mounted on a nose of the aircraft and are connected with the processing module; the inertial navigation system is connected with the GNSS satellite navigation receiving module; the starlight navigation device is mounted on the angle rotating mechanism; the light source is arranged at a set position on the ground; the inertial navigation system measures the angular velocity and acceleration of the aircraft; the GNSS satellite navigation receiving module receives a signal of a satellite; the angle rotating mechanism controls the horizontal direction of the optical axis of the starlight navigation device; the starlight navigation device captures a light source and images; and the processing module carries out navigation solution and Kalman filtering. According to the invention, rapid and effective alignment of the aircraft can be realized.

The utility model discloses a connecting mechanism, which relates to the technical field of distribution frames, and comprises a connecting unit comprising an L-shaped plate and a magnetic block fixedly connected to the side wall of the L-shaped plate; the clamping unit comprises a first fixing plate, an elastic assembly arranged in the middle of the first fixing plate, cylindrical groove rods symmetrically connected to the interior of the first fixing plate in a sliding mode, and abutting assemblies arranged at the tops of the cylindrical groove rods. And the dismounting unit comprises a baffle plate and cylindrical rods which are symmetrically and fixedly connected to the inner wall of the baffle plate. The utility model provides a device which can be quickly assembled and disassembled with the intelligent optical fiber distribution frame, double fixation is carried out through magnetic attraction between the connecting unit and the optical fiber frame and contact between the clamping unit and the optical fiber frame, and the fixing effect of the optical fiber distribution box and the optical fiber frame is improved. And the optical fiber distribution box is quickly mounted and dismounted by adjusting the dismounting unit, so that the mounting and dismounting time is shortened.

The utility model discloses a combined connection structure of a battery box, which comprises splicing seats clamped at two ends of a battery cell, two adjacent sides of each splicing seat are respectively provided with a first key strip, and the other two adjacent sides of each splicing seat are respectively provided with a first key slot matched with the first key strip; the adjacent splicing seats are mutually clamped through first key strips and first key grooves, the frame assembly comprises a first L-shaped frame body and a second L-shaped frame body, second key strips matched with the first key grooves are formed in the inner wall of the first L-shaped frame body, and second key grooves matched with the first key strips are formed in the inner wall of the second L-shaped frame body; the end parts of the first L-shaped frame body and the second L-shaped frame body are mutually clamped and locked; according to the lithium battery support structure, the problem that the upper ends of the battery cells are difficult to align and clamp in the holes of the support when the upper support is mounted and clamped due to the fact that the upper support and the lower support of the lithium battery support structure in the prior art are mostly of an integrated structure and the number of the battery cells is large is solved.

The application provides an infrared and visible light image fusion method based on text semantic consistency guidance, relates to the technical field of multi-modal image fusion, and comprises the following steps: fine-grained text semantics of infrared and visible light images are generated respectively and mapped to a unified embedding space; a cross-modal attention mechanism is used to bidirectionally compensate and enhance the text semantics, and a unified text semantic prior is constructed; a structure-intensity decoupled double-branch encoder is used to extract structure texture features of visible light and intensity saliency features of infrared respectively; under the guidance of the text semantic prior, explicit semantic consistency constraints and implicit semantic distribution consistency constraints are used to align the double-modal visual features in a shared semantic space; and finally, the text semantic prior is used as a global modulation signal to adaptively weight and fuse the aligned features and decode to generate a fusion image. The application effectively solves the problems of insufficient semantic modeling and poor fusion result consistency of existing methods.

The invention relates to the technical field of medical equipment, and discloses a data processing method for medical equipment, and the method comprises the steps: controlling the time of each part to be synchronized; inserting the collected data of each component into a synchronization mark; and aligning the collected data of each component through synchronous mark alignment. Through a synchronization mechanism, accurate synchronization marks can be inserted into all the components during data collection, so that in the subsequent data integration process, the collected data of different components can be accurately aligned according to the synchronization marks, the synchronism and consistency of the data are ensured, and limitation on hardware transmission is avoided. The invention further discloses a data processing device for the medical equipment and electronic equipment.

The invention discloses a target semantic segmentation method based on radar data and image data, and the method comprises the following steps: 1, carrying out the preprocessing of radar point cloud data: obtaining the radar point cloud data through a radar sensor, and carrying out the preprocessing of the radar point cloud data; 2, projecting the preprocessed point cloud data, and converting the point cloud data into a pseudo-depth map form; the method specifically comprises the following sub-steps: step 2-1, determining a projection plane; step 2-2, carrying out coordinate conversion; step 2-3, carrying out projection calculation; step 3, performing feature interaction: performing feature interaction on the radar data in the pseudo-depth map form and the input image data; and step 4, inputting the fused multi-modal features into a decoder network to generate a semantic segmentation result. By combining the advantages of the radar point cloud data and the image data, the invention provides a semantic segmentation method which can still keep high precision in severe environments such as low light, strong light, rain and fog and the like.

The large cactus pipe connecting tool comprises a base, a limiting groove is formed in the upper portion of the base, a lead screw is connected to the inner side of the limiting groove through a bearing, and a butt joint mechanism used for adjusting and aligning a large cactus pipe connector and a main pipe connector is arranged above the limiting groove. A first sliding groove and a second sliding groove are formed in the upper portion of the base, the first sliding groove and the second sliding groove communicate with each other, and pipe connectors of the cactus B and the cactus C can be adjusted to be aligned with a pipe connector of the pipe A through the second sliding groove. By means of the butt-joint mechanism and an inverted installation mode, the cactus B and the cactus C can be fully automatically abutted against pipe connectors on the two sides of the pipe A through the butt-joint mechanism, the cactus B and the cactus C do not need to be manually suspended and abutted by using a crane, the stability and the precision of whole cactus tailor-welding are improved, the welding efficiency is also improved, and the labor intensity of workers is reduced. And meanwhile, the safety is improved.

The present invention discloses a method and device for re-identifying a person in sketch based on reliable cross-modal subjective features, and relates to the technical field of character sketch image processing. The method and device for re-identifying a person in sketch based on reliable cross-modal subjective features mainly include: obtaining a training set, a validation set, and a test set according to the character RGB and sketch images; constructing a dual-path feature enhancement network using a residual network and training using the training set to obtain a trained dual-path feature enhancement network; and obtaining a character sketch retrieval result using the trained dual-path feature enhancement network according to the test set. Implementing the method and device for re-identifying a person in sketch based on reliable cross-modal subjective features provided by the present invention can improve the accuracy and robustness of cross-modal retrieval tasks.

This invention relates to the fields of biomedical imaging and spatial metabolomics, specifically to a method and system for registering spatial metabolomics images with HE-stained images. The method includes the following steps: acquiring two adjacent frozen sections of the same tissue sample; performing image preprocessing using the frozen sections; obtaining the extraction and matching results of key point features in the images; performing rigid image registration and obtaining a rigid registration result; performing non-rigid image registration and obtaining a non-rigid registration result; and obtaining the spatial location information of the spatial metabolic imaging using the non-rigid registration result. This invention, by combining advanced image processing technology and machine learning algorithms, automatically registers metabolic imaging images with HE-stained images, significantly improving registration accuracy and efficiency. It effectively solves the problems of large errors, low efficiency, and strong operational dependence in existing manual registration methods, meeting the needs of high-throughput, high-precision spatial metabolomics analysis and supporting large-scale data processing.

The application discloses a large language model training method based on mortise and tenon building drawing and text alignment, relates to the technical field of model training, and comprises the following steps: acquiring a mortise and tenon structure library and a conventional structure library; acquiring structure characteristic parameters of mortise and tenon structures by using a structure disassembly analysis method; acquiring structure conventional parameters of the mortise and tenon structures by using a conventional correlation method; constructing a mortise and tenon text training method, and training a large language model by using the mortise and tenon text training method; in the aspect of mortise and tenon recognition and discrimination, the application lacks an analysis method for point feature and geometric position constraint of mortise and tenon structures, cannot accurately recognize different mortise and tenon structures and judge the positions of the mortise and tenon structures, and thus problems of recognition confusion and text annotation errors and omissions exist when recognizing mortise and tenon in drawings.

The metal can edge pressing mechanism comprises a support, a hydraulic cylinder is installed in the middle of the support, a working table is arranged above the support, an outer cover is connected to the upper portion of the working table, a first telescopic sleeve rod is arranged above the working table, a limiting plate is installed on the inner wall of the outer cover, and a second telescopic sleeve rod is arranged on the outer cover. An electric telescopic rod is connected to the inner wall of the outer cover, a clamping block is installed at the tail end of the electric telescopic rod, a sliding block is connected to the outer end of the electric telescopic rod, and a driving wheel is meshed with the outer side of the sliding block; the device further comprises a driving unit, and the driving unit is arranged on the outer side of the first telescopic sleeve rod. And the edge pressing unit is arranged below the clamping block. By means of the edge pressing device and the reciprocating rolling mode, the problem that a gap is generated between the bottom plate metal sheet and the outer side of the lower portion of the metal can is solved, and the sealing performance of the bottom plate metal sheet combination pressed on the bottom of the metal can can be improved.

The present invention discloses a cross-project experience transfer method for intelligent customer service combined with transfer learning, S1, generating a source project feature set; S2, generating a source project optimization model including a shared feature layer; S3, globally optimizing the shared feature layer parameters in the source project optimization model through Bayesian optimization; S4, generating an optimized target project feature set; S5, building a target project model; S6, fine-tuning the target project model; S7, performing semantic consistency evaluation, response accuracy evaluation and business logic matching evaluation on the fine-tuned target project model on the target project data set; S8, deploying the target project model that has passed performance verification to the target project intelligent customer service system. The present invention realizes efficient deployment of a cross-project intelligent customer service system while significantly improving the migration efficiency and service quality of the model, and can be widely used in intelligent customer service scenarios in multiple industries.

The utility model discloses a handcart switch cabinet lifting adjusting mechanism which comprises a frame, a track plate is arranged on a table top of the frame, a circuit breaker body is slidably arranged on the track plate, a plurality of holes are formed in the table top, a sliding rod is arranged at the lower end of the track plate, and the circuit breaker body is arranged on the sliding rod. The sliding rods penetrate through the holes and extend into the frame, the lower ends of the multiple sliding rods are connected through a connecting plate, a lead screw rod is arranged in the frame, the lead screw rod is sleeved with a lead screw nut, and the lead screw nut is fixed to the connecting plate. The utility model relates to a handcart switch cabinet, and solves the problems that a base cannot be effectively aligned with a cavity of a cabinet body and a circuit breaker cannot be pushed in because the cabinet body of an existing handcart switch cabinet is placed in different places and sometimes the cabinet body is heightened or the ground is different in height.

The invention provides a model training method, a voice processing method and related equipment. The method comprises the following steps: acquiring a sample audio and a text annotation and an audio annotation corresponding to the sample audio; performing first training processing on the language processing model according to the sample audio and the audio annotation to obtain a first audio processing model; and performing second training processing on the first audio processing model according to the sample audio and the text label to obtain a target audio processing model. According to the embodiment of the invention, the stability and accuracy of a model processing result can be improved.

The invention discloses an image-text cross-modal Hash code extraction model establishment method and application thereof, and belongs to the field of cross-modal retrieval, and the method comprises the following steps: respectively extracting Hash codes of images and Hash codes of texts in a training set and a current training batch by using a Hash model, and establishing an image global graph, an image local graph, a text global graph and a text local graph, reconstructing text features by using the image local graph, and reconstructing image features by using the text global graph; establishing a loss function, and performing model training until the loss function converges; the loss function comprises intra-modal comparison loss which is used for shortening the distance between adjacent Hash codes in the global graph and shortening the distance between non-adjacent Hash codes; the inter-modal comparison loss comprises similarity loss between each node in the global graph and another modal Hash code matched with the adjacent node; and reconstructing the loss, including the loss between the reconstructed feature and the original feature. According to the method, the accuracy of image-text cross-modal retrieval can be improved.

This invention provides a traffic situation perception method, device, electronic device, and storage medium, belonging to the field of intelligent transportation technology. The method includes: acquiring multi-dimensional heterogeneous data, including human-dimensional data, vehicle-dimensional data, road-dimensional data, and cloud-dimensional data; acquiring the target scene categories corresponding to the multi-dimensional heterogeneous data and determining the scene priorities corresponding to the target scene categories; dynamically configuring perception resources according to the scene priorities to obtain configured perception resources, which are used to characterize sensor sampling frequency, sensor cooperation mode, and redundant device on / off status; and performing traffic situation perception based on the configured perception resources to obtain traffic situation perception results. The above solution can achieve real-time and accurate assessment of traffic situations, providing lane-level refined information for traffic management departments, supporting accurate decision-making, and effectively alleviating road network congestion and accident risks.

The utility model provides a double-layer corrugated paper production mechanism and a corrugated paper board production line applying the same, which comprises a corrugated roller, a corrugated peak a of a corrugated tooth of the corrugated roller is arranged to be a plane, so that a corrugated peak b of formed corrugated paper is also a plane; according to the corrugated roller, the corrugation peaks of the corrugation teeth are ground into planes, when two corrugating roller points are in butt joint in a tip-to-tip mode, the two corresponding corrugation peaks in the plane shape make parallel contact with each other to form face-to-face parallel attachment, and therefore the corrugation peaks of formed corrugated paper are also planes, and when the two pieces of corrugated paper are in butt joint and bonded, the corrugation peaks of the corrugated paper are not prone to deformation. According to the corrugated roller, the contact face is greatly increased through the planar ridges, bonding is very firm, based on the planar butt joint ridge structure, when two layers of corrugated rollers roll oppositely, the whole planar ridges are stressed integrally, stress is balanced and stable, it is ensured that the ridges of corrugated paper can be effectively aligned and cannot deviate, bonding pressure is enough, and the corrugated paper can be effectively bonded. And on the premise of effective alignment, firm bonding is further guaranteed.

This invention relates to a feature extraction model training and feature extraction method and apparatus, comprising: inputting a training image set into an initial model; extracting several local features from the training images by a local feature extraction layer; enhancing the several local features using an attention mechanism by a feature enhancement layer to obtain several enhanced local features; a global feature extraction layer aggregating features based on the several enhanced local features to obtain global features; and a classifier obtaining predicted classification parameters for the training images based on the global features. The model parameters of the initial model are adjusted according to the label parameters and predicted classification parameters of the training images to obtain a trained classification model. A feature extraction model for extracting global features is constructed based on the local feature extraction layer, feature enhancement layer, and global feature extraction layer in the classification model. This provides an end-to-end global feature extraction scheme, reducing algorithm complexity.

The present application relates to the technical field of computer data processing, and more particularly to a kind of exercise representation method and system based on knowledge concept and difficulty frequency alignment.Through steps: S1, construct knowledge concept representation learning based on causal attention mechanism;S2, establish difficulty estimator based on dynamic learning performance calibration;S3, knowledge concept and difficulty alignment based on semantic space consistency;S4, based on local window enhancement contrast learning strategy, the representation space is optimized.The present application can fully utilize learning path context, dynamic difficulty perception and local neighborhood information, improve the accuracy and stability of exercise representation.

The present invention belongs to the field of artificial intelligence technology, and provides a multi-round dialogue processing method and system based on RAG and knowledge graph. The method includes: according to the proprietary knowledge graph, the entities associated with the user's question are retrieved and identified to obtain the associated entities to build the reasoning link and convert it into the reasoning text sequence; the text, picture, and audio data related to the user's question are retrieved through the multimodal large model to obtain the relevant multimodal data; the relevant multimodal data, the current reasoning text sequence, the user's question and the generated answer text are integrated as the model input, and the multimodal large model is input to obtain the model answer; the relevant multimodal data is screened out, and the screened multimodal data is used as the context and the problem to be processed as the model input of the multimodal large model to obtain the answer text corresponding to the problem to be processed. The present invention can effectively align the user's intention with the answer text while reducing the interaction.

This invention relates to the field of network security technology and discloses a vulnerability graph alignment method based on text and graph embedding interaction, comprising the following steps: obtaining publicly available vulnerability data from the National Information Security Vulnerability Database and the National Information Security Vulnerability Sharing Platform; constructing a dataset for vulnerability entity alignment experiments; constructing a graph embedding representation model QuatAE based on the QuatE model; and introducing a new relation r. a Serving as a "connecting bridge" between two knowledge graphs, this vulnerability graph alignment method, based on text and graph embedding interaction, fully considers the characteristics of vulnerability graphs. It uses a unified model to comprehensively consider information at the entity, neighbor, and attribute levels, achieving good alignment results. Besides supporting vulnerability graphs, it is also compatible with entity alignment work for knowledge graphs in other domains because it incorporates information from various levels of the knowledge graph that affect knowledge alignment into the training process, demonstrating strong versatility.