121 results about "Co-training" patented technology

A system for real-time detection and mitigation of morphing malware in high-density edge networks, consisting of: a data acquisition unit configured to receive, normalize, and encode multimodal telemetry data streams originating from at least one of the following domains: network traffic, process behavior, system call sequences, binary instruction traces, and control flow graphs; the data acquisition unit is further configured to compute feature embeddings over sliding time windows and apply privacy-preserving redactions prior to storage; a generative neural processor that is operationally coupled to the data acquisition unit and configured to generate synthetic morphing malware variants by learning probabilistic transformations of previously observed malicious data representations, maintaining semantic functionality while varying structural and behavioral features; a discriminative neural processor trained adversarially with the generative neural processor, wherein the discriminative neural processor is configured to detect morphing malware by evaluating a probability distribution over multimodal telemetry embeddings and classifying anomalous process and flow behaviors in real time; a coordination processor that is communicatively connected to both the generative neural processor and the discriminative neural processor and is configured to orchestrate adversarial co-training, regulate detection thresholds, calculate reinforcement-based penalties for false negative results, and trigger countermeasures as soon as a detection confidence level exceeds a predefined adaptive threshold; a secure, system-integrated inference and enforcement unit configured to perform low-latency countermeasures at the network edge, including selective packet filtering, flow isolation, process interruption, or system microsegmentation, based on instructions from the coordinating processor; and a hardware-embedded security enclave that is embedded in the system and configured to store cryptographic keys, neural model parameters, and integrity affirmation data to ensure the confidentiality, authenticity, and immutability of model artifacts and policy configurations.

The invention discloses a cross-modal large model construction method and system based on track spatio-temporal characteristics, and belongs to the crossing field of artificial intelligence and dynamic spatio-temporal data processing, and the method comprises the steps: carrying out the sliding sampling and spatial distribution difference judgment through the semantic dynamic segmentation of multi-scale track spatio-temporal data, and generating spatio-temporal data blocks with consistent semantics; designing a space-time encoder of a hybrid architecture, extracting track time sequence association and spatial features, and unifying dimensions; constructing a text space-time fusion mechanism, and dynamically adapting cross-modal features by means of an anchor interface and gating fusion; a staged instruction fine tuning strategy is adopted, semantic alignment of space-time and text features is optimized firstly, then model top-layer parameters are trained cooperatively, complex scene adaptation is enhanced in combination with instruction difficulty progression and hard sample mining, and space-time constraint regular terms are introduced to guarantee output rationality. According to the method, high-precision cross-modal reasoning capability is provided for scenes such as track analysis and track prediction.

The invention belongs to the field of crossing of remote sensing visual questioning and answering and artificial intelligence, and particularly relates to a remote sensing visual questioning and answering method based on a large language model and a multi-level attention mechanism. Comprising the following steps: 1, constructing a MogaNet-based semantic segmentation network; 2, text coding is carried out through a large language model; 3, constructing a mixed attention guiding module; 4, constructing a bidirectional gating cross attention module; and 5, constructing and cooperatively training a remote sensing visual question-answer network. According to the MLaVQA framework provided by the invention, efficient and accurate visual question-answering processing of remote sensing images can be realized, when large-scale multi-modal remote sensing data is processed, complex semantic questions can be quickly answered, fine monitoring and intelligent management of earth resources are realized, the data utilization efficiency and decision support capability are remarkably improved, and the method is suitable for large-scale multi-modal remote sensing data processing. And technical support is provided for natural resource management, ecological protection and sustainable development.

The invention discloses a robust federated learning method for processing heterogeneous noise and non-independent identically distributed data, and belongs to the technical field of federated learning. The method provides a robust learning framework of two-stage client quality perception. The method comprises the following steps of: 1, constructing a category-level loss vector and clustering by using a Gaussian mixture model, and accurately dividing a clean and noise client set; stage 2, performing differential training: performing standard training on the clean client; dual-network cooperative training, dynamic sample screening and exchange, and a self-distillation and entropy regularization mechanism are introduced to a noise client, so that robust learning is realized; in the global aggregation stage, a distance sensing weighting strategy is further adopted to dynamically suppress the influence of a noise client; according to the method, original data does not need to be shared, the robustness and generalization performance of the federated learning model in the coexistence environment of heterogeneous noise and non-independent identically distributed data can be effectively improved, and the method has wide application value in the fields of medical images, financial risk control and the like.

The invention discloses an edge caching method for federal deep reinforcement learning based on semantic enhancement, and the method comprises the steps: obtaining request original data and a candidate content set, and carrying out the multi-modal feature extraction; calculating a cache utilization rate, request diversity and a dynamic semantic matching weight; calculating the semantic similarity between the request content and the candidate content, and performing adjustment in combination with a dynamic semantic matching weight to obtain a weighted semantic similarity; generating action probability distribution according to the state of the node, and fusing the action probability distribution with the semantic attention weight to obtain action preference distribution; action execution is selected according to the action preference distribution, and rewards are evaluated; and periodically executing federal cooperative training optimization, locally training a semantic perception strategy network by adopting an SAC algorithm, executing multi-factor dynamic weighted federal aggregation to update a global model, and distributing the global model to each node. According to the method, the edge cache hit rate and the semantic hit rate can be improved, the service response delay is reduced, and the robustness of heterogeneous nodes and a complex network environment is enhanced.

The invention relates to the technical field of intelligent operation of a power distribution network, and particularly discloses a power distribution network load prediction and scheduling system based on multi-modal data and semi-modal learning, and the system comprises the steps: dividing monitoring nodes into label nodes and label-free nodes through collecting the multi-modal data of the power distribution network in real time; constructing a power distribution network causal knowledge base containing load type and meteorological condition causal association by using the labeled node data; causal association is used as priori knowledge to configure a semi-supervised meta-learning model comprising a main prediction network and a knowledge query network; carrying out cooperative training by taking the labeled data as a supervision signal and taking the unlabeled data as a supplement; in the prediction stage, the initial load prediction of the main prediction network and the target causal association retrieved by the knowledge query network are combined, and physical rule conformity correction is achieved through calculation; and performing multi-level security check based on the corrected load prediction result, and automatically executing a hierarchical scheduling strategy according to the early warning level.

The invention discloses an urban traffic cooperative scheduling method and system based on a large model and multiple agents. The method comprises the steps that natural language task description is converted into a task semantic graph and a structured prompt; generating a plurality of strategy candidates by using a large language model, and performing language scoring and reasoning arbitration; based on the scoring result, selecting an optimal strategy for multi-agent execution; behavior execution data are collected and evaluated and fed back, and strategy closed-loop updating is achieved through parameter optimization and Prompt fine tuning; and a distributed task embedding mechanism and a lightweight migration module are combined, so that the multi-task adaptability and the training efficiency of the system are improved. The cooperative training system constructed by the invention has the capabilities of natural language interaction, strategy interpretability, behavior controllability and task migration, and is suitable for various complex urban tasks such as traffic jam dispersion, emergency response, signal lamp linkage and the like.

The invention relates to the field of artificial intelligence, can be applied to business system platforms of finance, medical health and the like, and discloses a classification method, device, equipment and medium based on cross-modal and dynamic distillation, and the method comprises the steps: collecting annotated image samples and unannotated text samples, and carrying out cross-modal bidirectional generation type enhancement processing and cross-modal sharing coding; performing teacher-student model cooperative training of dynamic distillation weight through annotated image samples and an extended data set; modal difference analysis is carried out according to the feature vectors of the unified semantic space, and distribution difference analysis is carried out on generated data and real data; and calculating a discriminant loss result according to each item of data so as to update model parameters, and obtaining a trained multi-modal classification model for classification processing. Small sample data distribution is effectively expanded through cross-modal generation of supplementary data and sharing of codes, the model convergence speed is increased through a dynamic knowledge distillation mechanism, the training efficiency is improved while the training effect is ensured, and the prediction classification accuracy is improved.

The invention discloses a multi-layer semantic perception, distillation and semi-supervised cooperative training target detection method, and the method specifically comprises the following steps: constructing a data set: extracting a first part of images from image data, marking the first part of images to construct a supervised target detection data set, and taking the remaining images as an unmarked image data set, the data volume of the unlabeled image data set is greater than that of the supervised target detection data set; teacher model optimization: performing supervision training on the teacher model on the supervised target detection data set; constructing a teacher model, and executing self-distillation training on the teacher model; pseudo labels are generated for the unlabeled images through a teacher model, and adaptive screening is carried out based on confidence distribution; mapping a pseudo label to a strong enhanced sample through enhanced geometric transformation, carrying out semi-supervised training by using the strong enhanced sample and the pseudo label, and introducing a feature layer distillation constraint at the same time; optimizing a student model; and outputting the target detection model obtained through training.

The invention belongs to the technical field of artificial intelligence and natural language processing, and particularly relates to a rule and regulation question and answer illusion suppression fine tuning method based on an anti-fact negative sample. According to the method, three types of anti-fact negative samples of error units, error thresholds and error versions are constructed, and cooperative training is carried out in combination with boundary maximization loss of evidence perception, numerical value consistency regularization, head rejection and temperature calibration and a controlled generation mechanism, so that the discrimination capability of the model on high-similarity error information is effectively improved; the accuracy of numerical answers is ensured, and the answers can be reliably rejected when evidences are insufficient or unreliable, so that the illusion phenomenon of a large language model in professional questions and answers is remarkably inhibited.

The invention discloses a cooperative training data generation method for an adjustment and optimization problem and electronic equipment, and the method comprises the following steps: carrying out the filtering processing of an input code description text, and obtaining input data; injecting the problem data into the input data to generate first adversarial sample data containing the problem data; gradient information and disturbance amplitude are calculated, and second adversarial sample data are obtained through adjustment and optimization; designing a shared encoder based on the second adversarial sample data to obtain problem data and normal code data features; carrying out separation processing on the two in a feature space; according to the separated features, the shared encoder is promoted to learn through feature comparison, and finally a model generated by code training data is obtained; and designing an evaluation index for judging whether a generation result of the model in the problem data injection scene is correct and stable, thereby completing model training of code training data generation. According to the method, generalization and robustness of the large language model in the process of processing actual problems are improved.

The invention discloses an English listening and speaking interaction intelligent training system based on a large language model, and relates to the technical field of English learning, and the training system comprises a user capability graph module, an interaction feedback module, a personalized learning path generation module, a scene simulation module, an interaction rhythm control module and a community cooperative training module. According to the method, the pragmatic ability, the cross-language communication ability and the linguistic ability are jointly incorporated into the training content, and the adaptive training content is generated in combination with a real cross-language communication scene, so that a user synchronously masters expression specifications and culture adaptation key points in different scenes in the training process; the effect that training content and real communication requirements are deeply matched is achieved, the user is helped to accurately adapt to a native English scene and real phrase habits of people, the requirements of an actual application scene are met, and substantial improvement from basic language knowledge mastering to real communication ability is achieved.

This application discloses a hierarchical obstacle avoidance method and system based on collaborative training of the PPO-TD3 algorithm. This application utilizes a dynamically adaptive hierarchical decision-making system, namely a collaborative architecture of the PPO and TD3 algorithms, to achieve intelligent obstacle avoidance in dynamic environments through a hierarchical reinforcement learning framework. In simple environments, only the high-frequency control network of the underlying TD3 algorithm is run to improve response speed, while in complex scenarios, the complete hierarchical architecture is activated to ensure decision quality. Simultaneously, a semantic instruction interface is innovatively designed, enabling the abstract instructions output by the high-level PPO decision-making layer policy network to be intelligently parsed into executable actions by the underlying TD3 execution layer control network in conjunction with real-time physical constraints. Combined with a hybrid training framework integrating simulation training, human demonstration, and real-world learning, and a resilient safety mechanism based on dynamic risk assessment, this achieves synergistic optimization of safety and operational efficiency in complex dynamic environments.

The present application relates to the technical field of computer deep learning, and relates to a device fault diagnosis network model training method based on multi-source domain knowledge joint migration, which comprises the following steps: constructing a training sample set, a verification set and a test set; training a pre-constructed device fault diagnosis network model by using the training sample set, optimizing the parameters of the device fault diagnosis network model, and obtaining a trained device fault diagnosis network model; verifying the trained device fault diagnosis network model by using the verification set, evaluating the diagnostic performance of the trained device fault diagnosis network model, and selecting an optimal device fault diagnosis network model; and testing the optimal device fault diagnosis network model by using the test set, and evaluating the performance of the optimal device fault diagnosis network model. The present application uses the multi-source domain collaborative training of the model by using the collected multi-source domain data set containing a small amount of labels composed of a plurality of labeled source domain data sets and unlabeled source domain data sets, learns the field invariant features in the fault samples, and migrates the shared diagnostic knowledge from the labeled data and the unlabeled data to the target field, so that the trained device fault diagnosis network model can complete the cross-domain fault diagnosis task without needing to obtain the target samples in advance.

The application discloses an emotion recognition method based on online cross-modal knowledge distillation, comprising the following steps: acquiring electroencephalogram and electrocardiogram original signals and windowing and cutting; constructing electroencephalogram and electrocardiogram student models, extracting intermediate features from each modal data segment through an encoder, and obtaining non-normalized prediction output through a classifier; constructing a teacher probability distribution through a joint encoder fusion; introducing adaptive contrast loss to align the cross-modal intermediate features, introducing distillation loss to constrain the prediction probability distribution of each modal to align with the teacher probability distribution; synchronously optimizing new student model parameters through online collaborative training; and performing actual inference prediction based on the student model after training. The application combines double modal signals to make up for the defects of single modal information, excavates the complementarity of modes, realizes dynamic generation of teacher supervision signals and real-time collaborative learning of modes through online distillation, does not increase test calculation overhead, effectively improves the recognition accuracy, model robustness and generalization ability, and has good application prospect.

The invention provides a test case generation method and system based on multi-modal machine learning and dynamic optimization, and the method comprises the steps: carrying out the cross-modal feature association of a demand semantic text, a source code instruction stream and a historical defect event record in a software test scene, generating a multi-modal association feature space, inputting a dynamic attention alignment network, and carrying out the multi-modal feature association. Weight distribution is carried out on the business rule semantic dependency relationship, the code execution path dependency relationship and the defect propagation link association relationship, a joint representation vector is generated, double-task cooperative training of test case generation and defect positioning prediction is executed on the joint representation vector, feature extraction weights of double tasks are adjusted to generate a test case generation parameter set, and the test case generation parameter set is subjected to defect positioning prediction. Inputting the boundary test scene into a generator of the generative adversarial network, strengthening feature expression of the boundary test scene, generating a boundary test case set, performing priority ranking on the boundary test case set to calculate case coverage weight, and generating a target test case set. According to the invention, the effectiveness and comprehensiveness of the test case and the utilization efficiency of test resources are improved.

The invention relates to the technical field of artificial intelligence, and provides a semi-supervised image recognition method utilizing open set sample enhancement, and the method comprises the steps: building a cooperative training system integrating a multi-scale feature extraction backbone network and a triple classifier through preparing a data set containing closed set category label samples and unlabeled samples containing open set samples; the multi-scale feature extraction backbone network extracts multi-scale fusion features from the unlabeled samples, and inputs the multi-scale fusion features into a triple classifier; based on the rejection probability output by the negative discrimination classifier and the similarity intensity output by the class centroid similarity classifier, generating a dynamic confidence score according to a reverse coupling relationship, and selecting an unlabeled sample of which the confidence score exceeds a dynamic threshold value as a candidate sample for enhancing the closed set recognition capability; for candidate samples, pseudo labels are generated by a closed set main classifier, and parameter optimization is carried out by combining the following three types of signals; and dynamically updating network parameters through weighted fusion of Ls, Lsu and Lcu until convergence. According to the technical scheme, high-value open set samples are accurately screened, and the closed set image recognition performance is remarkably enhanced.

The application discloses a kind of based on semi-supervised learning and improved YOLOv8's fissure intelligent identification method and system, belong to fissure identification technical field.First, by semi-supervised labeling strategy constructs high-quality fissure image dataset, and adopts three-stage process to realize the standardization conversion and visual verification of labeled data;Second, based on improved YOLOv8n-seg light weight instance segmentation network, fusion multi-scale feature pyramid and triple data enhancement strategy, through three branch detection head realizes the collaborative training and optimization of classification, positioning and segmentation;Finally, via confidence filtering, non-maximum suppression, mask optimization and coordinate system restoration system post-processing, output structured recognition results containing boundary box and pixel-level mask.The application has significant advantages in improving small target detection capability, boundary positioning accuracy and model generalization performance, and is suitable for geological engineering safety monitoring and disaster warning scene.

The application provides a deep multi-view clustering method and system based on collaborative training, relates to the fields of data mining and machine learning, and specifically includes the following steps: obtaining multi-views to be clustered, and pre-training a deep self-encoding model for each view; calculating an affinity matrix of each view, setting an initial weight of each view and an initial dynamic learning factor between the views; performing iterative formal training on the deep self-encoding model of each view, updating the weight of each view and the dynamic learning factor between the views until a preset iteration stopping condition is met, and outputting final hidden layer features and clustering centers; and obtaining a clustering result based on the final hidden layer features and the clustering centers; the application designs a deep multi-view weighted graph embedding clustering algorithm based on dynamic collaborative training, adopts a dynamic collaborative training idea for multi-view data, deeply mines complementary information between the multi-views, and improves the clustering effect for multi-view data sets.

The invention discloses an antibody language model training method and device and an antibody sequence prediction method and device, and the method comprises the steps: obtaining a training sample set, and enabling training samples in the training sample set to be generated through a pre-constructed antibody multivariate coding vocabulary; performing structured corpus construction processing on training samples in the training sample set to obtain a training corpus data set; and according to the training corpus data set, performing dual-task cooperative training processing on a language model, and generating a pre-training antibody language model for antibody design. The data processing capability of the antibody language model can be improved, and the method can be widely applied to the technical field of artificial intelligence.

The application discloses a kind of based on open source big model's poetry beginner multi-task cooperative training method, comprising: by knowledge isolation module, the semantic classification of user input text is carried out, when identifying as poetry field content, the preset knowledge base call of open source big model is blocked, and enters incremental learning procedure, when identifying as daily communication content, response is generated directly based on open source big model, and the scheme can identify whether user is learning or daily communication by six module cooperative linkage, and the preset knowledge base call of open source big model is blocked as appropriate, solve the defect that existing model learns before knowing, guarantee progressive learning, improve the effect of poetry learning, realize that child is as teacher to teach AI doll this beginner learning knowledge, simulate real learning process, let child incarnate teacher, doll incarnate beginner, with output instead of input, child can more profound repetition and understanding knowledge in the process of repetition.

The application belongs to the technical field of computer vision, and discloses a zero sample picture classification method, system, device and medium based on collaborative learning, which comprises the following steps: acquiring a data set; training a picture encoder and a semantic encoder; extracting picture encoding and semantic encoding; dividing two sub data sets; training two classifiers by using picture encoding and semantic encoding of the two sub data sets respectively; predicting test data by using the two classifiers, and dividing the test data into the two sub data sets according to the ranking of scores; judging whether the sub data sets are updated; repeatedly dividing the two sub data sets to train the classifiers until the sub data sets are no longer updated; and finally predicting by using the two classifiers finally obtained. The application trains two classifiers by using picture features and semantic features respectively in a collaborative training mode, fuses the complementary knowledge in the picture features and the semantic features in a mixed two-classifier data mode, and improves the accuracy of class classification.

A centrifugal circulating pump group cooperative training method based on federated learning belongs to the technical field of artificial intelligence and industrial Internet of Things, and constructs a federated learning system composed of a central aggregation server and a plurality of edge clients, and realizes safe transmission of model parameters through gRPC encryption communication. The edge client carries out model training based on pump group operation data collected locally, and Gaussian noise is added by adopting a differential privacy technology to protect data privacy; the central server performs secure aggregation on the encrypted model weight update, generates an enhanced global model and distributes the enhanced global model to each client; and each node further realizes model personalized adaptation through local fine tuning, and realizes low-delay deployment by using model quantification and an RKNN inference engine. According to the method, on the premise of ensuring that the data is not delocalized, the generalization ability, the control precision and the system robustness of the model in the heterogeneous pipe network environment are effectively improved.

The invention discloses a federated learning-based multi-modal semantic unified fusion retrieval method. The method comprises the following steps of: locally preprocessing multi-modal data by a plurality of participants; cooperatively training and aggregating model parameters under a federated learning framework to generate a global model; and after the global model is utilized to encode each modal feature, the modal features are input to a high-order cross-modal fusion encoder. The encoder constructs a tensor network and performs contraction operation on the tensor network so as to explicitly model high-order semantic association among different modes and generate a unified semantic vector. And finally, calculating the similarity between the query vector and the unified semantic vector in the database according to the query request, and generating a retrieval result. The method aims at solving the problems that in the prior art, fusion depth is insufficient, and high-order semantics are difficult to process, high-order complex association among multi-modal data can be effectively captured, data privacy is protected through federal learning, and meanwhile semantic representation precision and retrieval accuracy are improved.

The invention discloses a data confusion method in collaborative training and collaborative reasoning of a multi-layer perceptron, and relates to the field of data confusion processing, and the method comprises the steps: in a forward propagation stage of training the multi-layer perceptron, for an lth network layer in the multi-layer perceptron, carrying out the forward propagation of the multi-layer perceptron; generating a first output and a second output of the lth network layer based on the first parameter and the second parameter of the lth network layer and the first input and the second input; and for the lth network layer in the multi-layer perceptron, generating a first gradient and a second gradient of the lth network layer based on the first feedback and the second feedback of the lth network layer, and updating a first parameter and a second parameter of the lth network layer according to the first gradient and the second gradient of the lth network layer. And the demand balance of three impossible triangles of security, light weight and high efficiency in data confusion is considered.

The invention discloses an emotion recognition method based on online cross-modal knowledge distillation. The emotion recognition method comprises the steps that electroencephalogram and electrocardio original signals are obtained and subjected to window segmentation; constructing an electroencephalogram and electrocardio student model, extracting intermediate features from each modal data segment through an encoder, and obtaining non-normalized prediction output through a classifier; teacher probability distribution is constructed through joint encoder fusion; self-adaptive comparison loss alignment cross-modal intermediate features are introduced, and distillation loss is introduced to constrain prediction probability distribution of each modal to align to teacher probability distribution; new student model parameters are optimized synchronously through online cooperative training; and performing actual reasoning prediction based on a student model after training. According to the method, bimodal signals are combined to make up single-modal information defects, modal complementarity is mined, teacher supervision signal dynamic generation and modal real-time collaborative learning are realized through online distillation, test calculation overhead is not increased, identification precision, model robustness and generalization ability are effectively improved, and the application prospect is good.

The invention discloses an education practical training scoring and optimizing method based on multi-agent procedural cooperative training, and relates to the technical field of intelligent education and multi-agent cooperation. According to the method, a collaborative system composed of a practical training guidance agent, a special evaluation agent group, a confrontation feedback agent and a global coordination agent is constructed. The method comprises the following steps: performing sub-task decomposition on an education practical training task, and performing procedural and multi-dimensional evaluation on the practical training task in combination with multi-modal data such as codes, voices and videos; a confrontation feedback mechanism is introduced, credibility correction is performed on a special evaluation result, and scoring robustness is improved; and through local continuity constraint and global consistency optimization, dynamic balance between the sub-task score and the overall target is realized. According to the method, the real-time performance, traceability and iterative optimization of education practical training scoring can be realized, the objectivity, fairness and generalization ability of evaluation results are improved, and the method is suitable for multi-type online education and occupational practical training scenes.

The invention discloses a fine granularity identification method based on adaptive threshold cooperative training, which comprises the following steps: carrying out determinacy-based subdivision on a prediction consistent sample set to obtain a high-determinacy sample set and a low-determinacy sample set; carrying out confidence-based subdivision on the predicted inconsistent sample set to obtain a correctable sample set and an irrelevant sample set, and carrying out label correction on the correctable sample set; performing cross updating on the two neural networks by using the high-determinacy sample set and the correctable sample set; and updating mean teacher network parameters and the like. According to the method, samples which are inconsistent in prediction but high in correction confidence coefficient are corrected by adopting pseudo tags of the mean teacher network, and information of difficult samples is fully utilized instead of being simply discarded. According to the method, an adaptive threshold mechanism based on the statistical characteristics of the current batch of samples is provided, the certainty threshold and the confidence threshold are automatically adjusted in the training process, and the robustness and universality of the method are improved.

This invention discloses a machine learning-based method and system for predicting groundwater level changes, relating to the field of hydrogeology. The method includes: standardizing multi-source time-series data to obtain a standardized multivariate time-series data matrix; constructing a supervised learning sample set; inputting the groundwater level sequence from the supervised learning sample set into a physically-guided variational mode decomposition network; decomposing the groundwater level sequence into K intrinsic mode component sequences and a residual term sequence using a loss function with physical-driven consistency constraints; for each of the K intrinsic mode component sequences, dynamically assembling a differentiable simulator from a library of differentiable simplified physical simulators, and co-training these simulators with the goal of approximating each intrinsic mode component sequence and reconstructing the original water level sequence as a whole, resulting in K fully trained assembled differentiable simulators. This invention generates reliable, visualized prediction results through multi-simulator collaborative extrapolation and uncertainty quantification.

The invention discloses a data-free model acquisition method considering accuracy and robustness, and belongs to the technical field of artificial intelligence safety. The method comprises the following steps: S1, introducing a generator jointly guided by classification, inter-class and intra-class diversity loss, and synthesizing a high-quality high-entropy sample to detect a target model decision boundary; s2, constructing a main and auxiliary substitution double-model collaborative framework, wherein an auxiliary substitution model carries out self-exploration by generating an adversarial sample so as to learn robustness knowledge; s3, the main substitution model receives dual supervision of a hard tag from the target model and a soft tag from the auxiliary model by using the generated high-entropy sample, and performs cooperative training; and S4, outputting the finally obtained main substitution model with accuracy and robustness. According to the method, through a task decoupling double-model mechanism, under the condition that original training data does not need to be accessed, the accuracy and robustness of the target model are cooperatively obtained, and the adversarial robustness of the obtained model is remarkably improved.