78 results about "Mutual learning" patented technology

The invention relates to the technical field of federated learning, in particular to a dynamic federated mutual learning method and system for balancing individuation and generalization, and the method specifically comprises the following steps: each client carries out the preprocessing of data to be processed of a model, and carries out the strong enhancement and weak enhancement processing; inputting the data subjected to strong enhancement processing into a shared model, inputting the data subjected to weak enhancement processing into a private model, and performing iterative training on the two models; related parameters of the shared model after each round of iterative training and a difference item between two model parameters are uploaded to a federation server; the federated server adopts a multi-dimensional adaptive aggregation strategy to obtain an updated global model, and returns the updated global model to each client to replace the shared model in the next round of training; and finally generating a generalization result and a personalized result. According to the method, the private-shared model architecture is constructed, and dynamic federated mutual learning is carried out in combination with the federated server, so that balance and collaborative improvement of individuation and generalization performance can be realized.

The invention discloses a sparse trajectory pattern learning method based on mutual learning semantic distillation, and particularly relates to the field of trajectory analysis. Comprising the following steps: acquiring spatial characteristics of a road section to which a track belongs and spatial-temporal characteristics of each track point, and determining an embedding vector of each track point according to the spatial-temporal characteristics and spatial characteristics of the track points; constructing track features based on semantics and semantic features based on tracks; determining a track-based semantic vector according to the semantic-based track features, and determining a semantic enhanced track vector according to the track-based semantic features; fusing and interacting the semantic vector based on the track and the semantic enhanced track vector to obtain an embedded vector of the track; and inputting the embedding vector of the track, the initial occurrence time, the position of the interest point and the cue word into a pre-trained LLM text encoder to obtain a track mode. On the basis of the method, mode missing detection caused by data sparsity can be avoided, and then the accuracy of trajectory analysis is improved.

The invention discloses a knowledge distillation method for target detection. The knowledge distillation method comprises two major designs including a dynamic teacher-student mutual learning mechanism and interpretable feature decoupling. The RGB detector is used as a teacher guidance event detector to learn static semantic knowledge in a conventional illumination scene, and the event detector is used as a teacher guidance RGB detector to learn dynamic robust features in an unconventional illumination scene, so that dual-mode collaborative optimization is realized; and the original teacher and student feature space is decoupled into three parts of mode universality, mode specificity and mode irrelevance, and effective knowledge migration based on the mode universality feature is carried out between the two modes by constructing bidirectional distillation loss. The objective of the invention is to enable an RGB detector and an event detector to have a more accurate target recognition effect.

The invention provides a cloud-edge collaborative federation mutual learning training method for electric power AR operation and maintenance information decision, and the method is applied to an edge server. The method comprises the steps that a joint gradient of edge servers is generated based on soft label prediction distribution of other edge servers in a collaboration group, computing power levels of the edge servers, bidirectional KL divergence loss and task loss and uploaded to a cloud end, the joint gradient is used for the cloud end to calculate a global gradient, and the global gradient is used for updating parameters of a global model; extracting sample data features of the target power equipment according to the initial light quantum model to obtain sample local features and uploading the sample local features to a cloud end, wherein the sample local features are used for the cloud end to calculate sample feature residual errors; obtaining sample complete features based on the feature residual and the local sample features, and inputting the sample complete features into the initial light quantum model to obtain local output distribution; and iteratively updating the initial light quantum model based on the optimization target to obtain a target light quantum model. According to the method, the model training efficiency is improved while the data privacy security is guaranteed.

The invention discloses an intelligent identification and digital management system for cable and accessory identifiers. The system comprises an identifier image acquisition module, a character feature extraction module, a character identification processing module, a parameter correlation analysis module, a digital modeling module and a data storage management module. According to the system, a multi-model integration framework combining a PaddleOCR character recognition algorithm and a mutual learning distillation strategy is adopted, so that the identification recognition stability in a complex environment is improved; constructing a parameter incidence relation graph by means of a graph convolutional neural network, and verifying parameter integrity and consistency; a digital modeling module is used for forming a digital twinborn body with embedded parameters, and space structure and parameter information fusion is achieved. The full-process automatic processing reduces manual intervention, improves the management efficiency, provides data support for the full-life-cycle management of cables and accessories, promotes the intelligent and refined upgrading of the operation and maintenance management of a power system, and is suitable for large-scale and high-complexity cable and accessory management.

The invention provides a heterogeneous model knowledge migration method based on federated learning. The method comprises the following steps: a data collaborative preparation stage; screening local data, performing multi-dimensional labeling, and performing cooperative training and federal modeling preparation on a plurality of participants; a privacy initialization stage; privacy protection processing is carried out on local data, and local model parameters are initialized; a self-learning stage; each participant completes localized updating of the model parameters by using the local data set; a mutual learning stage; the server extracts valuable knowledge fragments from the knowledge representations uploaded by the participants through the super network for mixing, and updates model parameters of the participants; a server-side super network module is updated and iteratively cycled; and the server side aggregates the valuable knowledge fragments, updates the super network parameters and then returns to the step S3, and loop iteration is carried out until the model converges. The problems that computing resources of participants are limited, generalization is insufficient due to model heterogeneity, and a knowledge migration technology cannot adapt to a heterogeneous model architecture are solved.

The invention provides a farmland plot extraction method based on an agricultural field plot instance segmentation framework, and relates to the technical field of remote sensing image processing, and the method comprises the following steps: obtaining a remote sensing image of a farmland plot, and carrying out the preprocessing of the remote sensing image; constructing an agricultural field instance segmentation framework network model adapted to the unsupervised field; using the training set to train the agricultural field instance segmentation framework network model adapted to the unsupervised field; and performing farmland plot extraction on a target domain remote sensing image to be identified through the trained unsupervised field adaptive agricultural field instance segmentation framework network model. According to the method, the UDA-FT network model is constructed, and a teacher-cross-domain student model architecture and a consistency mutual learning strategy are adopted, so that the domain offset problem of cross-region and cross-sensor farmland plot extraction is effectively solved, and the precision and robustness of farmland plot extraction are improved.

The invention discloses a domain adaptive remote sensing image target detection method and system, and the method comprises the steps: introducing a teacher-student learning normal form based on a Deformable DETR architecture, and constructing a feature alignment mechanism under the guidance of a frequency domain; the method comprises the following steps: firstly, carrying out model preheating training by utilizing source domain data, and fusing CNN and Transform encoder features in combination with a frequency domain filter; further constructing category prototypes of the source domain and the target domain, and applying prototype confrontation loss and inter-prototype comparison loss; in the pseudo label generation process, a dynamic threshold strategy guided by a prototype is introduced, high and low confidence degree samples of a target domain are divided in a self-adaptive mode by combining the sample confidence degree and the prototype similarity, and a category pseudo label judgment standard is dynamically updated; through a two-stage training mechanism, target domain detection learning is guided by using a high-confidence-coefficient pseudo tag in a teacher-student model mutual learning stage, frequency domain modeling and a prototype modulation mechanism are fully combined, and the robustness and generalization ability of cross-domain target detection are effectively enhanced. And the performance in adaptive tasks in the field of remote sensing images is excellent.

The invention provides a low-illumination rainy day image enhancement method based on DEPMD-Net. Through heterogeneous double teachers and multi-stage distillation, polarized rain-light attention and implicit representation are introduced, separation and fusion of raindrop high-frequency information and illumination low-frequency information are realized, and an image with balanced brightness and without rain interference is generated. The method comprises the following steps: 1) synthesizing low-illumination rainy day data based on Rain100L and LOL-v1; 2) pre-training a brightness expert and a rain removal expert, and obtaining prior through bidirectional KL mutual learning; 3) utilizing brightness-rain mask statistics to construct joint priori, and adopting non-zero mean anchoring to simulate night rain distribution in diffusion noise scheduling; 4) proposing PiDNet-X to recover features through cooperation of a double-branch structure and rain-light attention, 5) performing staged distillation through a soft label and feature prompt, obtaining lightweight Student-C / F through a Born-Again strategy, and realizing night rain scene self-correction by double students in a reasoning stage through moving average and mutual learning.

The application discloses a kind of cloud edge collaborative scene based on deep mutual learning's cigarette packaging defect detection method, based on the cigarette packaging image sample of different edge end is constructed defect detection dataset, and the global model of different size defect detection model and cloud end is built, utilize defect detection dataset and alternately circulate and execute mutual learning strategy and the global parameter updating strategy under cloud edge collaborative scene training model, fixedly trained each edge end small model parameter and obtain reconstruction error threshold value;Utilize edge end small model processing to be measured packaging appearance image and obtain reconstructed image, according to the comparison relationship of reconstruction error and reconstruction error threshold value of reconstructed image, determine whether there is appearance defect.The application combines the framework of cloud edge collaboration, fuses the information between different cigarette factories and cigarette production lines, breaks the information silos in industrial production, effectively solves the problem of data distribution difference caused by different imaging equipment.

The invention discloses a key negotiation synchronization method and system for a complex satellite network, and the core of the key negotiation synchronization method is that chaotic system synchronization and neural network mutual learning synchronization are organically fused, and a dual synchronization architecture is constructed; specifically, a local field of each hidden unit of the TPM is transformed and then input into improved Logistic chaotic mapping, and then output bits of the hidden units are determined after output of the chaotic mapping is transformed through an arcsine function, so that a bidirectional feedback coupling relation is established between state updating of a neural network and evolution of a chaotic system. According to the design, the fast synchronization characteristic of the chaotic system and the mutual learning capability of the neural network are mutually promoted, the synchronization convergence speed of key negotiation is remarkably increased, meanwhile, the attack resistance of the system is enhanced by utilizing the unpredictability of chaos, and finally, on the premise that the safety is guaranteed, the key negotiation efficiency is improved. And the efficiency and the reliability of the key negotiation process of the complex satellite network are effectively improved.

The application discloses a fundus photograph classification system based on deep mutual learning. The application fuses two kinds of anomaly detection algorithms of Ganomaly and FastFlow, constructs a detection model which is adaptive and robust to various eye diseases, effectively solves the problem that artificial intelligence cannot be used for intelligent diagnosis of a disease due to the lack of abnormal samples of the disease in practical application, improves the universality and accuracy of the disease detection model, saves the precious time of doctors, and improves the diagnosis efficiency of patients. In addition, the detection method can be applied to a fundus camera, and the device plays a great role in the initial screening stage of eye diseases. With slight cost, the device solves the problem that early eye diseases of patients are difficult to be found due to objective reasons such as insufficient medical resources and lack of experienced doctors, and truly helps community and primary hospitals to realize early discovery, early diagnosis and early treatment, so that the tragedy of blindness caused by delayed treatment due to late discovery is avoided.

The invention relates to the field of image processing, and provides a self-supervised image denoising method and system for a multi-angle sequence, a better denoising effect is obtained through pixel rearrangement processing and noise spatial correlation destroying, and compared with traditional mask convolution or cavity convolution, through a blind spot network based on a displacement-cutting strategy, the denoising effect is improved, and the denoising efficiency is improved. The weight arrangement of the convolution kernel does not need to be changed, the standard convolution performance is reserved, the structure is simple and efficient, the hardware requirement is low, the convolution kernel is allowed to directly feel the center pixel through the non-blind-spot network to obtain complete context information, collaborative optimization is carried out according to the mutual learning optimization sharing loss function, and the method is suitable for large-scale popularization and application. According to the self-supervised image denoising method, the denoising image is output only through the non-blind spot network with less parameter quantity after optimization training, noise is effectively suppressed, the structural details of the image can be kept, higher structural fidelity is achieved, and the denoising accuracy and efficiency of the self-supervised image denoising method are improved.

Systems and methods of providing intelligent energy-efficiency performance coaching. Energy efficiency performance is measured by a performance score calculated in association with various efficiency metrics. The performance score may be provided as feedback to the operator and may further be used by an operator reward system and / or gamification system. An intelligent energy-efficiency coach responds to variability in operator behaviors by using tunable variables to bias weighting constants assigned to the efficiency metrics to normalize calculations of performance scores. The intelligent energy-efficiency coach may use reinforcement machine learning techniques to calculate tunable variables and adjust performance scores on an ad hoc basis. A mutual learning process may be performed, where the operator may learn energy saving behaviors from feedback provided by the intelligent energy-efficiency coach to improve energy efficiency and the intelligent energy-efficiency coach adjusts tunable variables to help the operator improve their energy efficiency performance.

The invention discloses an equipment maintenance robot group for mutual learning of maintenance experience and a working method thereof. The system comprises a cloud server, at least three maintenance robots, an edge computing node and a distributed sensor network. The cloud server stores a global maintenance experience library, and a block chain-database hybrid architecture is adopted to ensure that data cannot be tampered and is efficiently stored. The maintenance robot integrates a multi-modal data acquisition module, a federated learning module, a heterogeneous communication module, a maintenance execution module and an energy management module. The edge computing node is responsible for local experience verification, conflict resolution and model aggregation, and cloud load is reduced. And the distributed sensor network feeds back the equipment state in real time to assist in fault pre-judgment. Through the steps of multi-modal data acquisition and preprocessing, local experience generation, encryption and uploading, edge verification and aggregation, global collaborative optimization, dynamic task allocation and execution and the like, robot group cross-regional maintenance experience mutual learning is realized, and the intelligent level and overall efficiency of equipment maintenance are improved.

The invention discloses a semi-supervised pedestrian re-identification method based on multistage clustering. The method comprises the following steps: acquiring video data; and inputting the video data into a pedestrian re-identification model to obtain a pedestrian identification result. Wherein the pedestrian re-identification model is obtained according to the following steps: training a feature extraction network by using a labeled training set; for unlabeled unsupervised data, extracting corresponding output features by using the trained feature extraction network; performing single-track clustering, single-video clustering and cross-video clustering on the output features to obtain label information, and further constructing an automatic labeling pedestrian training set; carrying out distillation training on a teacher-student mutual learning framework based on the automatic marking pedestrian training set and supervised data, wherein the teacher-student mutual learning framework comprises a teacher model and a student model; and taking the student model subjected to distillation training as the pedestrian re-identification model. According to the invention, the data annotation cost is reduced, and the generalization ability of the pedestrian recognition model is improved.

This invention discloses a remote USB sharing system, including a sharing unit, a seller login center unit, and a communication unit. It also discloses a remote USB sharing method comprising six steps. The remote USB sharing system proposed in this invention features a simple architecture, low development cost, ease of operation, and convenient real-time communication, facilitating mutual learning and exchange among colleagues. It is particularly beneficial for novice sales personnel, allowing for quick onboarding, effectively reducing training costs, and significantly improving operational efficiency. For buyers, the order placement and login processes are also relatively simple, and the system includes communication and consultation functions, greatly enhancing the buyer's experience when purchasing remote USB dongles. This promotes its widespread adoption. The remote USB sharing method proposed in this invention has simple and reasonable steps, enabling real-time communication between buyers and sellers, which facilitates order fulfillment, significantly improves sales performance, and ultimately enhances corporate profitability.

The invention discloses a pedestrian re-identification method based on an identity reservation cross-view mutual learning framework, and belongs to the technical field of image identification, and the method comprises the following steps: obtaining a plurality of weakly enhanced pedestrian image samples and strongly enhanced pedestrian image samples; constructing a first convolutional neural network and a second convolutional neural network, and correspondingly generating a first time average model and a second time average model; constructing a first positive and negative sample set and a second positive and negative sample set based on the hard pseudo tag of the convolutional neural network; forming a first positive comparison target set and a second positive comparison target set based on the soft pseudo labels of the convolutional neural network; constructing a comparison loss function, a classification loss function and a triple loss function to form a total loss function; obtaining a trained collaborative neural network; and performing pedestrian re-identification on the newly acquired pedestrian image to obtain a re-identification result of the target pedestrian. According to the method, the problem that pedestrian re-identification is easily influenced by background noise and false label errors is solved.

The application discloses a cross-subject electroencephalogram signal classification method based on online test time domain adaptation, and steps of the method comprise the following steps: 1, pre-processing original EEG data, including removing noise, segmenting, extracting time-frequency features by using short-time Fourier transform, and obtaining source domain data and target domain data; 2, constructing a source model, a student model and a teacher model based on a CNN network, and training the source model by input data to obtain a pre-training source model; 3, initializing the student model and the teacher model by using the pre-training source model; 4, online optimizing the student model and the teacher model based on a mutual learning strategy on target data flow and realizing classification of electroencephalogram signals. The application can realize rapid classification of electroencephalogram signals under the condition of protecting the privacy of patients, so that the real-time demand of the classification system of electroencephalogram signals in an actual scene can be met.

The present invention discloses a method for re-identifying occluded pedestrians, comprising: constructing a multi-task learning network, including a full-body pedestrian identity re-identification branch and an occluded pedestrian identity re-identification branch; constructing a loss function corresponding to each of the two branches; L h and L o ; The mutual learning strategy is used to train the multi-task learning network. The CNN feature expression parts of the two branches share weights, that is, first use L h The whole body pedestrian identity re-identification branch is trained, and the weight of the CNN feature expression part obtained by training is shared with the CNN feature expression part of the occluded pedestrian identity re-identification branch, and then used L o The occluded pedestrian identity re-identification branch is trained, and the weights of the CNN feature expression part obtained through training are shared with the CNN feature expression part of the full-body pedestrian identity re-identification branch. The training of the two branches is performed alternately; occluded pedestrian re-identification is performed based on the trained multi-task learning network to obtain recognition results, thereby improving the accuracy of pedestrian re-identification.

The invention belongs to the field of pattern recognition, and discloses a fusion model for underwater target recognition, which comprises the steps of signal preprocessing; extracting features in the audio signals to form a data set of the invention; dividing the data set into a training set, a verification set and a test set according to a proportion of 8: 1: 1; a fusion model is built based on an improved MobileViT backbone network and a branch network based on a multi-spectrum attention mechanism to solve the problem that a traditional deep learning model cannot give consideration to both recognition precision and recognition efficiency, and a training set and a verification set are input into the fusion model for training; inputting the test set into the trained network to identify the ship type; the fusion model comprises a backbone network, a branch network and deep mutual learning. The backbone network adopts a decoupling full-connection attention mechanism to replace a multi-head self-attention mechanism in the MobileViT to reduce the complexity of the model. The branch network adopts a multi-spectrum attention mechanism to provide more detail features for the trunk network. The deep mutual learning can fuse the feature information of the backbone network and the branch network. The fusion model can give consideration to both recognition precision and recognition efficiency.

The present invention relates to a cross-domain target detection method and system, a device, and a storage medium. The method comprises the following steps: converting source-domain image data into new source-domain image data on the basis of CycleGAN; constructing a framework based on mutual learning between a student model and a teacher model, and updating the student model and the teacher model on the basis of the new source-domain image data and a pre-constructed target detection model; and using the teacher model meeting a preset condition as a final model to perform target detection on target-domain image data, to obtain a target detection result. The present invention can be widely used in the technical fields of intelligent algorithms, machine learning, and computer vision.

The invention provides an elevator management system, electronic equipment and a readable storage medium, and the elevator management system is designed with modules for mutual learning in production inspection, field debugging, operation self-adaption stages and data maintenance stages of elevators in parallel connection and group elevators. After inspection, debugging, operation self-adaption, maintenance updating and fault maintenance of one elevator are completed, relevant information is shared to other number machines through the learning module, and after other number machines learn, learning updating or auxiliary personnel updating is completed by themselves according to different modules. Through the mutual learning module mechanism, the inspection, debugging, operation self-adaption and maintenance period of the whole elevator system is shortened, the investment of manpower and material resources is reduced, meanwhile, the overall operation efficiency of the system is improved, it is ensured that elevator parameters with the necessity of parameter consistency in the elevator system are kept consistent, and the system reliability is improved. Potential safety hazards possibly caused by inconsistent parameters of the elevator room are effectively eliminated.

The application discloses an image segmentation model training method and device, equipment and a storage medium, which utilizes a small amount of first sample image data with labels and a large amount of second sample image data without labels to input into a first segmentation network and a second segmentation network to respectively perform prediction, obtain respective segmentation results of the first segmentation network and the second segmentation network, and respectively perform supervised training on the first segmentation network and the second segmentation network based on a loss function constructed based on the segmentation results and the labels. Meanwhile, a pseudo-label semi-supervised loss constructed based on the segmentation result of the first segmentation network and the segmentation result of the second segmentation network is used to guide mutual learning of the first segmentation network and the second segmentation network. The application can reduce label sample data required for training a model, complete training of the model by using a small amount of sample data with labels, and improve robustness of the model by using mutual guidance and learning of the two segmentation networks.

This invention proposes a fingerprint ridge distance estimation method and system based on deep learning, relating to the field of biometric recognition technology. It comprehensively considers spatial and frequency domain information to construct a multi-scale fingerprint ridge distance estimation model, and further optimizes the model through mutual learning to make the measurement estimation more accurate. The specific scheme includes: cropping a training fingerprint image with labeled ridge distances into multiple labeled image patches; grouping labeled image patches with close distances between two labeled points to construct a training dataset; inputting the labeled image patches in pairs into two fingerprint ridge distance estimation models according to the groups; using structural similarity loss between two labeled image patches in the same group to promote mutual learning between the two models; cropping the fingerprint image to be estimated into multiple image patches, and inputting a single image patch into the final fingerprint ridge distance estimation model; calculating the fingerprint ridge distance of the fingerprint image based on the predicted ridge distance estimate of a single image patch sample.

The application is suitable for the technical field of model pruning, and provides a deep neural network pruning method based on dynamic contrastive mask and knowledge distillation.The application proposes a dynamic pruning framework based on dynamic contrastive mask learning and mutual knowledge distillation to compress a deep neural network model.Dynamic contrastive mask learning aggregates similar masks through contrastive learning and distinguishes them in a feature space to generate a binary mask;combined with a sample complexity adaptive factor, a sample subnetwork is adaptively pruned without manually setting a pruning threshold.Mutual knowledge distillation realizes knowledge transfer and feature enhancement through mutual learning between subnetworks to alleviate feature information loss in the pruning process.The framework can reduce model size and computational overhead, improve inference speed, maintain high accuracy, and improve model performance and robustness, and is suitable for scenarios with limited computing resources, can realize efficient channel pruning under the premise of ensuring deep neural network accuracy, and further improve model generalization ability.

The steps of the EEG emotion recognition system based on contrast learning and implicit emotion regulation mechanism are as follows: first, the preprocessed EEG feature matrix is divided into left and right brain two-dimensional EEG feature matrices according to the left and right electrode distribution respectively; the obtained and a randomly initialized adjacency matrix are input into a dynamic connection EEG representation extraction module to obtain left and right brain shallow emotion representations and left and right brain deep emotion representations respectively, and an emotion classification loss is calculated. Then, the obtained and are input into an automatic reverse regulation module to calculate a contrast loss; the obtained and are input into a brain lateralization mutual learning module to calculate a KL loss. The total loss obtained from and is used to constrain the system, and EEG emotion recognition network parameters γ are obtained. Finally, the EEG data to be tested is input into the EEG emotion recognition network, and the final emotion recognition result is obtained using γ. The present application further improves the EEG emotion recognition accuracy from the perspective of biological mechanism.

The present invention relates to the field of artificial intelligence technology, and specifically discloses a mutual learning method and system between multi-view medical images and text reports. By constructing a mutual learning model and dividing the learning process of the mutual learning model into an image reconstruction task, a report reconstruction task, and a multi-view alignment task, the image reconstruction task obtains the feature representation of the multi-view medical image, and the report reconstruction task obtains the feature representation of the text report and cross-modally fuses it with the image feature representation to obtain a reconstructed text report. The integration of multimodal reconstruction tasks enables the model to learn richer and more detailed feature representations, thereby improving its robustness to lost or damaged data. In addition, the exchange-based multimodal fusion method in cross-modal text reconstruction aims to fully integrate visual features and enrich the semantic representation of domain-specific knowledge. Through pre-training and joint training, the model performance is optimized so that more accurate and comprehensive diagnostic results can be obtained after applying the mutual learning model.

The invention belongs to the technical field of artificial intelligence and block chains, and relates to a low-orbit satellite constellation-oriented model collaborative credible training system and method. The system comprises a main chain and a plurality of fragments, the main chain is used for task release and model mutual learning, and the fragments are used for model distributed training; the method comprises the following steps: fragmenting the constellation according to the orbit plane distribution characteristics of the low-orbit satellite constellation and completing configuration; the nodes in the fragments are trained, evaluated and aggregated according to identities allocated by the intelligent contracts, aggregation model parameters are generated, node reputation updating is carried out, after the main chain obtains the aggregation model parameters uploaded by the fragments, a plurality of nodes with sufficient energy are selected to complete a model mutual learning process, and then optimized model mutual learning results are fed back to the fragments. And a final training result is obtained through multi-round iteration. According to the method, the interference of the malicious edge satellite nodes in the low-orbit satellite constellation on the model training process can be effectively inhibited.

The application discloses an open knowledge base-oriented concept dependency relationship identification method, medium and equipment, relates to the technical field of natural language processing, and mainly comprises the following steps: training a concept representation learning model by using a concept description text set and a contrast learning target function on a pre-trained language model; obtaining a concept dependency relationship identification model by using a deep mutual learning method according to the concept representation learning model; and identifying a concept dependency relationship by using the concept dependency relationship identification model. The open knowledge base-oriented concept dependency relationship identification method, medium and equipment provided by the application can improve the robustness and computational efficiency of concept dependency relationship identification.