348 results about "Eigenvector computation" patented technology

The invention relates to an intelligent interview scoring system capable of explaining decisions based on a large language model, and the system comprises a multi-mode resume analysis and feature coding unit, a resume feature adaptive matching unit, an interactive scoring and knowledge enhancement unit, and an answer quality evaluation unit. Text, image and audio features are extracted through a cross-modal attention mechanism of a multi-modal large language model, resume features are encoded into dynamic word vectors, and entity-level feature vectors are extracted; the post description text is encoded into a demand feature vector by a resume feature adaptive matching unit; calculating semantic similarity between the resume entity feature vector and the demand feature vector; the interactive scoring and knowledge enhancement unit dynamically retrieves knowledge fragments to generate a preliminary evaluation report containing a scoring basis; and the answer quality evaluation unit fuses the information density, the fluency and the integrating degree to generate a final score. And the whole-process intelligence from demand analysis to final decision making is realized.

An assessment apparatus and method for objective pain assessment are provided, wherein the apparatus includes a processor comprising a frequency-domain transformation module, a frequency-band segmentation module, a first assessment submodule, and a second assessment submodule. The frequency-domain transformation module generates a global time-frequency feature matrix, which is segmented by the frequency-band segmentation module in a frequency domain into five frequency bands associated with pain perception. The first assessment submodule extracts last time step features and an adjacency matrix from the time-frequency feature matrix and generates a first feature vector representing global association patterns among electrodes used for acquiring EEG signals. The second assessment submodule generates a second feature vector with local spatiotemporal dynamic features of EEG signals, concatenates it with the first feature vector to form a fused feature vector, computes its class probability distribution, normalizes it, and generates an objective pain quantification indicator corresponding to the EEG signals.

The invention provides a weak supervision video anomaly detection method and system based on potential energy field damping dynamics. The method comprises the following steps: inputting a video into an anomaly detection model to obtain an original video feature sequence; calculating fluctuation potential energy by using the original video feature sequence and further obtaining a global inertia proxy vector; calculating interaction potential energy through the global inertia proxy vector and the original video feature vector; nonlinear self-adaptive damping force is generated through interaction potential energy; applying a nonlinear adaptive damping force to the original video feature vector to obtain a purified dynamic feature vector; constructing a loss function by using the purified dynamic feature vector, and training the model to obtain an optimized model; and inputting the video into the optimized model to obtain a final frame-level anomaly detection result. According to the method, anomaly detection is reconstructed from a classification problem to a signal decoupling and energy dissipation problem in a physical system, and the physical nature of a depolarization mechanism is explained from the theoretical level.

The invention discloses an adaptive evaluation method for the health degree of an electrolytic cell, and the method comprises the following steps: collecting the multi-dimensional operation parameters of the electrolytic cell in real time, and carrying out the preprocessing of the collected time series data, so as to construct a training sample with a time window; extracting a multi-scale time sequence feature from the training sample to form a feature vector; inputting the feature vector into a weight adjustment network, and outputting a dynamic weight vector; weighting the feature vector by using the dynamic weight vector to generate a weighted feature vector; inputting the weighted feature vector into a performance prediction model, and outputting a short-term performance prediction value of the electrolytic cell at a future moment; after the corresponding real performance value is obtained, calculating a prediction error of the short-term performance prediction value, and constructing a reinforcement learning reward signal based on the prediction error; updating a strategy of the weight adjustment network through a reinforcement learning algorithm by utilizing a reward signal, thereby optimizing dynamic weight vector generation at a subsequent moment; based on the dynamic weight vector and the feature vector at the current moment, a comprehensive health degree index of the electrolytic bath is obtained through calculation; according to the method, main factors influencing the equipment health degree in different stages are intuitively revealed, and a basis is provided for operation and maintenance decision making.

The invention discloses a multi-dimensional training method and device for a support vector machine, electronic equipment and a computer readable storage medium, and the method comprises the steps: carrying out the discretization of a training sample data set, and obtaining a discretized data set, the discretized data set comprises a plurality of different attributes, and each attribute corresponds to a plurality of feature vectors; calculating a classification contribution parameter of each attribute feature vector to obtain a plurality of classification contribution parameters; performing data mapping on the plurality of classification contribution degrees by using a kernel function to obtain a target function; and optimizing and training the objective function by using a gradient descent algorithm to obtain a support vector machine model. According to the method, different dimension data are mapped through the kernel function, the data gain weight can be determined, the linear separable effect of the mapped dimension data can be improved, and then the classification and analysis capability of the SVM can be improved.

The invention relates to the technical field of new energy power system instability risk assessment, in particular to a new energy power system frequency instability risk assessment method based on a heterogeneous graph attention network. The method comprises the following steps: constructing a wind power-photovoltaic heterogeneous graph model according to a power grid topology; performing combined sampling on different operation modes and anticipated disturbances, and determining corresponding input feature vectors; calculating a frequency stability index value label of the sample set; expanding the training set through an active learning iteration process, and carrying out model training; and inputting the collected operation data into the trained heterogeneous graph attention model, outputting a frequency stability index value, and evaluating the system frequency instability risk in combination with the risk matrix. By adopting the frequency instability risk assessment method for the new energy power system based on the heterogeneous graph attention network, the problem of low efficiency of risk assessment in a high-dimensional uncertain scene is solved, and the heterogeneous graph attention network can reflect the influence of different types of devices at different positions and disturbance types on the dynamic frequency of the system; and the accuracy of frequency instability risk assessment is improved.

The invention provides a network security malicious code binary search method and system, and relates to the technical field of data processing, and the method comprises the steps: calculating the multi-dimensional structural similarity between a to-be-detected binary file and each standard binary reference file based on a function level control flow semantic feature vector; determining a finally matched standard binary reference file according to the multi-dimensional structural similarity; performing fine-grained difference analysis on the binary file to be detected and the finally matched standard binary reference file to obtain a fine-grained difference analysis result; and on the basis of a fine-grained difference analysis result, in combination with sensitive data operation behavior feature detection, judging whether the to-be-detected binary file is a maliciously tampered version, and positioning a malicious code injection point. According to the method, the defects that in the prior art, dependence on fixed features is too high, and compilation optimization is sensitive are overcome.

The invention discloses a chromatographic data optimization processing method for a liquid chromatograph, and relates to the technical field, and the method comprises the following steps: collecting an original signal flow, carrying out preliminary peak detection, and generating an original peak list; extracting modal exclusive characteristics of each chromatographic peak; converting the modal exclusive features into a unified multi-dimensional semantic feature vector; calculating a time interval weight, and calculating a semantic relevancy weight based on the multi-dimensional semantic feature vector; correcting the time interval weight, generating a network edge weight, and constructing a peak correlation topology network by taking the multi-dimensional semantic feature vector as a node; identifying node clusters connected with high edge weights, and aggregating the node clusters into candidate compound entities; tracking a node signal intensity change track, and adding a conflict mark; conflict resolution arbitration is carried out on the candidate compound entities with the added conflict marks, dominant detector evidence is output, and an analysis report is generated. According to the method, the problems of poor data collaboration of multiple detectors, inaccurate peak identification and association and isolated data processing flow are effectively solved.

The invention discloses an image classification method based on multi-manifold joint metric learning, and the method comprises the steps: 1) data preprocessing: constructing Grassmann features and SPD features; 2) constructing a double-flow kernel matrix, including constructing a Grassmann manifold kernel matrix and an SPD manifold kernel matrix, and respectively calculating a Grassmann test kernel matrix and an SPD test kernel matrix; 3) implementing multi-manifold joint metric learning, and extracting kernel feature vectors from the kernel matrix to calculate an intra-class scatter matrix and an inter-class scatter matrix; solving the projection matrix, sorting the feature vectors according to the sizes of the feature values, and selecting the feature vectors corresponding to the first plurality of maximum feature values to form the projection matrix; the method comprises the steps of (1) extracting different ground features, (2) carrying out feature fusion, completing extraction and implementing sample mapping, and (3) constructing a sample-level fusion convolutional neural network and completing classification of different ground features of an original image.The method belongs to the technical field of image processing and machine learning, and the classification accuracy can be remarkably improved.

The invention belongs to the field of seat control, and particularly relates to a seat control method based on multi-modal data fusion in order to solve the technical problems that in the prior art, model interpretability is poor, and adjustment logic and physiological comfort requirements of a human body are inconsistent, and the seat control method comprises the following steps that S1, after a first feature vector and a second feature vector are mapped to a quaternion space, the first feature vector and the second feature vector are mapped to the quaternion space; generating an initial interaction tensor by calculating the Hamiltonian product of the two; s2, using KL divergence between prior distribution and empirical distribution as a regular term for constraint, and resolving a human body state vector; s3, inputting the human body state vector into the generative network to obtain a fusion feature vector; and S4, based on the fusion feature vector, calculating a control signal used for adjusting the seat surface posture of the seat, the protrusion amount of the waist supporting structure and the clamping angle of the side wing wrapping structure. According to the invention, the analysis and fusion of the vehicle movement trend are realized, so that the seat control can fully foresee and cope with the influence of the vehicle movement on the driver and passengers.

The invention relates to the technical field of lithium battery SOH estimation methods, in particular to a lithium battery SOH estimation method based on an EIS ensemble learning algorithm. Comprising the following steps: S1, collecting battery performance data; s2, collecting corresponding electrochemical impedance spectroscopy data through an EIS method; s3, obtaining feature data through ICA, DVA and DTV methods, performing normalization processing on the feature data and the data obtained in the S2, and merging the data into a feature vector; s4, calculating a capacity fading rate CAR; according to the capacity fading rate, allocating to different algorithms to carry out SOH estimation, and when the capacity fading rate is less than or equal to 10%, selecting an ELM algorithm to calculate an SOH estimation value; when the capacity fading rate is greater than 10% and less than or equal to 30%, selecting a CNN architecture to calculate an SOH estimated value; when the capacity fading rate is greater than 30%, selecting an SVM algorithm to calculate an SOH estimated value; and S5, displaying and storing the predicted SOC result. Compared with the prior art, the optimal estimation algorithm is dynamically switched based on the capacity fading rate, the full life cycle estimation precision is improved, and the calculation complexity is remarkably reduced.

The invention relates to the technical field of data mining and product demand analysis, and discloses a user demand mining method based on feature fusion density perception, a terminal and a medium. The method comprises the following steps: collecting product rejection data, extracting a semantic feature vector from a preprocessed data text, obtaining a user behavior sequence by using a session identifier in meta-information to generate a behavior feature vector, and generating a situation feature vector based on time and environment attributes in the meta-information; calculating the attention weight of each dimension feature, and performing weighted fusion on the semantic feature vector, the behavior feature vector and the situation feature vector by using the attention weight to generate a multi-dimensional feature representation vector; clustering the multi-dimensional feature representation vectors to obtain a plurality of clusters; and feature portrait construction and value evaluation are carried out on the cluster, a demand description document is generated, and optimization algorithm parameters are fed back according to a verification result. According to the method, the efficiency and the quality of mining potential user requirements from neglected rejection data can be remarkably improved.

The invention relates to the technical field of applet interface design, and discloses an adaptive design method and system based on an applet interface. The method comprises the following steps: acquiring screen parameters, image resolution, pixel density and screen size of the terminal equipment, and user operation habit data, such as touch frequency and interface staying duration; and generating equipment adaptive feature vectors according to the data so as to describe terminal equipment display characteristics and user interaction preferences. And constructing an interface element dynamic layout model, and calculating the relative position and the size proportion of the interface component through the model according to the equipment adaptive feature vector. And monitoring a transverse and vertical screen switching event of the terminal equipment in real time, and once the event is monitored, triggering the interface element dynamic layout model to recalculate. And sending the recalculated interface component layout data to a rendering engine to complete the adaptive updating of the applet interface. According to the method, the user experience of the applet on different devices can be effectively improved, and the adaptability and flexibility of the applet to various terminal devices are enhanced.

The invention discloses a multi-modal learning data conjoint analysis method and system, a medium and a product, and relates to the field of multi-modal learning analysis. Comprising the following steps: acquiring synchronously acquired multi-modal data and extracting a feature vector; and calculating a confidence score of each modal feature vector and a semantic conflict coefficient between the modal feature vectors. When the semantic conflict coefficient is greater than a preset conflict threshold value, determining the modal feature vector with the maximum confidence score as a final state feature vector; and when the semantic conflict coefficient is smaller than or equal to the threshold value, calculating a fusion weight according to the confidence score, the semantic conflict coefficient and the task priority parameter, performing weighted fusion on each modal feature vector to obtain a final state feature vector, and generating an evaluation result according to the final state feature vector. According to the invention, through a decision-making mechanism that the optimal information source is selected in high conflict and context adaptive fusion is carried out in low conflict, the problem of inaccurate evaluation caused by forced fusion of contradictory signals is solved, and the accuracy and robustness of learning state evaluation in a complex scene are improved.

The invention discloses a voiceprint-based early diagnosis method for a mechanical fault of a transformer. The method comprises the following steps: S1, collecting a voiceprint signal during operation of the transformer; s2, processing the voiceprint signal; s3, constructing a comprehensive feature vector; s4, calculating the matching degree between the feature vector and the reference feature in the fault sample library; s5, outputting a diagnosis result; according to the invention, by effectively separating the mechanical fault features of the transformer from environmental interference, the problem that a single-channel signal is easily polluted by noise is solved, a purer data source is provided for subsequent feature extraction, and the identifiability of early fault features is improved; by comprehensively capturing fault feature differences, the characterization capability of comprehensive feature vectors on early mechanical faults is remarkably improved; through hierarchical diagnosis and in combination with an early warning probability, common faults are rapidly matched by using a fault sample library, and special faults are accurately identified by virtue of deep learning, so that the accuracy and generalization ability of early diagnosis of mechanical faults are remarkably improved.

The invention discloses a communication data analysis system and method for network security, and relates to the technical field of computer internet. Time sequence features, protocol semantic features and interactive topology features of communication session historical data are extracted based on network flow data; the method comprises the following steps of: acquiring a time sequence feature, a protocol semantic feature and an interactive topology feature, fusing the time sequence feature, the protocol semantic feature and the interactive topology feature into a unified high-dimensional feature vector, acquiring a communication behavior record and a communication behavior dynamic feature vector of network equipment in a communication network, calculating a digital feature of a coupling relationship evaluation value, and when a certain communication session occurs, judging whether the communication session occurs or not. The method comprises the following steps of: calculating real-time coupling relationship evaluation values among network equipment, quantifying the difference degree of the real-time coupling relationship evaluation values through digital characteristics, accumulating the coupling relationship evaluation values in the process of performing a certain communication session, calculating the total anomaly degree of the communication session, and calculating the abnormal degree of the communication session. The method aims at solving the problems that advanced persistent threats are difficult to effectively recognize, feature expression is insufficient and the perceptual ability is weak in the prior art.

The invention discloses a deep learning time sequence alignment method and system for a multi-view asynchronous sequence image, and the method comprises the following steps: carrying out the image screening of a multi-view asynchronous collected angiography sequence, and constructing a short sequence composed of three frames of continuous images; the method comprises the following steps of: extracting a backbone network by adopting an OfficientNet-B0 as a feature extraction backbone network, removing a classification head of the feature extraction backbone network, replacing the classification head with identical mapping, outputting convolutional features, and performing network optimization in combination with a soft dynamic regularization alignment loss function (soft DTW); extracting a feature vector describing the heart cycle progress through a network; calculating the cosine similarity of the feature vectors among different view angles, constructing a similarity matrix, and searching an optimal alignment path in the similarity matrix by using a soft dynamic time warping algorithm; and outputting a time sequence alignment frame and a corresponding similarity score according to the optimal alignment path. According to the method, the multi-view asynchronous sequence image can be effectively processed, time sequence alignment is realized, and a reliable alignment basis is provided for cardiovascular image analysis.

The invention relates to the technical field of digital content management, in particular to a cultural creative work display management method and system based on artificial intelligence, and the method comprises the following steps: S1, obtaining the real-time behavior data of a target user in a current session; s2, extracting a user interest vector based on the real-time behavior data; s3, obtaining digital assets of the to-be-displayed cultural and creative works, and extracting multi-dimensional works feature vectors of the works; s4, calculating a matching degree weight between the user interest vector and the multi-dimensional work feature vector; s5, carrying out weight sorting on each feature item in the multi-dimensional work feature vector; and S6, calling corresponding display elements from a preset display resource library, and combining to generate a personalized display interface. According to the method and the device, the personalized display interface can be generated based on the matching result of the real-time interest of the user and the multi-dimensional features of the works, so that the display content is dynamically matched with the preference of the user.

The invention provides a hyperspectral data optimization method and device, equipment and a storage medium, and relates to the technical field of image processing. The method comprises the following steps: acquiring to-be-processed hyperspectral data, and performing multi-scale feature extraction on the to-be-processed hyperspectral data to obtain a multi-scale feature vector; calculating the fuzzy correlation degree between different wavebands of the to-be-processed hyperspectral data, performing waveband screening according to the fuzzy correlation degree to obtain a reserved waveband, and performing dimensionality reduction on the to-be-processed hyperspectral data according to the reserved waveband to obtain dimensionality-reduced hyperspectral data; dividing the dimension-reduced hyperspectral data into a plurality of local areas, and performing adaptive denoising on each local area according to local data features of each local area to obtain denoised hyperspectral data; and performing data reconstruction according to the multi-scale feature vector and the denoised hyperspectral data to obtain optimized hyperspectral data. According to the method, the hyperspectral data can be optimized more efficiently and accurately, and the quality and the application value of the hyperspectral data are improved.

The invention discloses a complication prediction method and system based on an SHAP-random forest, and relates to the technical field of medical data mining. The method comprises the following steps: acquiring preoperative CT images and puncture path planning data, and calculating risk factors; performing numerical value standardization, nonlinear transformation and interactive feature generation on the risk factors to obtain feature vectors, calculating mutual information scores of feature values in the feature vectors, if the mutual information scores are greater than an experience threshold, retaining the feature values, and after traversal is finished, obtaining updated feature vectors; on the basis of a random forest model, taking the updated feature vector as an input value, and calculating a complication probability; quantizing the contribution degree of the characteristic value based on a Shapley value; according to clinical indexes, the risk threshold is dynamically corrected, the complication risk level is divided, and complication prediction is completed, the problems that a static threshold ignores the blood coagulation state difference of a patient and a black box model cannot provide a decision basis are solved, and the complication misjudgment probability is reduced.

The invention discloses a multi-source data fusion-based real-time state monitoring method for power transformation and distribution electric equipment, and the method comprises the steps: extracting a dynamic change trend of equipment operation according to a unified multi-source data set, marking the dynamic change trend as an abnormal sequence when the dynamic change trend exceeds a preset threshold value, and obtaining an abnormal point distribution data set; calculating a trend prediction value according to the fused feature vector, performing judgment according to a trend preset value, and generating an early warning signal; according to correlation of interaction influence parameters extracted from the early warning signal, inputting correlation data into a neural network model for optimization, and obtaining an optimized interaction influence model; performing iterative updating on the multi-source data set according to the optimized interaction influence model, determining that the updated data set reflects the real state of the equipment, and obtaining a comprehensive state evaluation result; and generating a fault prediction report according to the comprehensive state evaluation result, and activating a real-time monitoring adjustment mechanism to maintain stable operation of the power grid if the potential fault probability in the fault prediction report is judged to be higher than a threshold value.

The invention discloses a zero-sample industrial anomaly detection method and system based on mixed Gaussian expert geometric constraints, and the method comprises the steps: constructing a mixed Gaussian expert prompt distribution model, and generating a mixed Gaussian expert distributed prompt set adaptive to image semantics; implementing prompt semantic geometric constraint and fusion sampling, adjusting the geometric position of an expert prompt semantic center, fusing the weight of each semantic expert through expert gating, and generating a final text prompt based on a re-parameterization sampling strategy; respectively extracting feature vectors of a final text prompt and an input image by using a text encoder and an image encoder; and calculating the cosine similarity between the text feature vector and the image feature vector as a scoring basis for anomaly detection. The method has excellent performance in the aspect of improving the zero sample industrial anomaly detection precision, and has important engineering application value for improving the intelligent detection level of the quality in the manufacturing process.

The invention relates to the field of computer intelligent control, and discloses an equipment control instruction generation method and system fusing a large and small model and a rule engine, and the method comprises the steps: collecting industrial equipment multi-mode sensor data, extracting features, constructing task feature vectors, calculating task emergency degree and complexity scores, obtaining task classification labels, and obtaining a task classification result; the task is routed to a fast system or a slow system to generate a candidate control instruction set, after rule engine physical boundary verification and digital twin sandbox effect rehearsal, an optimal instruction is selected and converted into an industrial protocol to be written into a programmable logic controller to be executed, and a decision link is recorded for model iterative optimization; the real-time performance, the safety and the optimization effect of an industrial equipment control instruction are ensured by constructing a fast and slow system cooperative processing architecture and a multi-level verification mechanism.

The invention relates to the technical field of data mining analysis, in particular to a petrochemical operation state evaluation method and system based on big data, and the method comprises the following steps: collecting a sensor signal to generate standard time sequence monitoring data, constructing a high-dimensional space-time correlation map based on physical topology, extracting a global state feature vector through a map attention network, and carrying out the analysis of the global state feature vector. And calculating a feature distance and mapping to generate a petrochemical operation state evaluation result. According to the method, a high-dimensional space-time correlation graph based on physical topology and data correlation is constructed, discretely distributed sensor signals are mapped into graph structure nodes with logic connection, and hidden nonlinear coupling characteristics among multi-source parameters are deeply mined by using a graph attention network model; compared with the prior art, the method overcomes the defect that a single-point monitoring mode splits process continuity, achieves the precise quantification of the performance degradation trend of equipment under a complex working condition in combination with the dynamic evaluation logic based on the feature space distance, and effectively improves the global consistency and accuracy of petrochemical operation state evaluation.

An SAR and multispectral fusion method based on low-rank adaptive fine tuning comprises the steps that a low-rank adaptive layer is constructed, a model is initialized, in initialization, weight pre-training is conducted on newly-added low-rank matrix parameters through a remote sensing visual language model RemoteCLIP, and pre-training parameters are frozen; the SAR image and the multispectral image are subjected to wave band grouping through a wave band grouping and prompt guiding mechanism based on physical attributes, the SAR image is divided into a VV group, a VH group and a PolSAR group through grouping, and multiple spectrums are divided into an RGB group, a VRE group, an NIR group and an SWIR group through grouping; respectively extracting image features and semantic prompt text features corresponding to each wave band group through an image encoder and a text encoder to obtain corresponding image feature vectors and text feature vectors; calculating the similarity between the image features and the text features to obtain predicted image category labels; calculating classification loss and comparison loss, carrying out network training, and repeating training until the model converges; and the model obtained by training is used for reasoning.

The invention relates to the field of post matching, and particularly provides a person and post matching method and system and a storage medium, and the method comprises the following steps: obtaining target post information and multi-dimensional data of candidate persons, the multi-dimensional data comprising dynamic behavior data with timestamps and capability evaluation data; based on the timestamp of the multi-dimensional data, constructing a space-time weight matrix by adopting a preset time decay function, and performing weighting processing on the multi-dimensional data to generate timeliness talent features; inputting the timeliness talent features and the target post information into a preset deep learning model to generate talent feature vectors and post feature vectors; calculating the similarity between the talent feature vector and the post feature vector to obtain an initial matching degree; and based on the initial matching degree and a preset post performance prediction model, performing dynamic feedback optimization on a matching result to generate a final man-post matching scheme. The technical problem that in the prior art, data are static and cannot reflect dynamic changes is solved.

The invention relates to the technical field of fire fighting detection, and discloses a fire fighting access occupation detection system based on machine learning. The system comprises a fire fighting access multi-modal data acquisition module, a multi-dimensional feature fusion module, a channel occupation feature learning module and a dynamic monitoring module. The multi-modal data acquisition module acquires a real-time monitoring image frame sequence, and synchronously acquires depth sensor and infrared thermal imaging data; a multi-dimensional feature fusion module extracts image space features, and constructs a multi-modal feature fusion sequence in combination with depth and infrared data; a channel occupancy feature learning module carries out modeling to generate a spatial topological feature map and extracts a static occupancy feature vector; the dynamic monitoring module obtains environment vibration and people flow density data, calculates dynamic occupation features by combining the static feature vectors, and outputs real-time dynamic change feature vectors. The system can comprehensively capture static and dynamic occupancy states of the channel, and is suitable for complex environment supervision.

The invention relates to the technical field of industrial process control, and discloses a plastic product large-scale production multi-parameter cooperative control system and method. A deviation calculation module; an adaptive compensation module; and a set value updating module. The method comprises the following steps: injecting a probe signal in a preset non-sensitive stage, synchronously acquiring a response signal, and calculating a process response feature vector; calculating the deviation between the measurement vector and a preset reference vector to generate a measurement deviation vector; determining a target compensation vector through deviation attribution in combination with prediction drift, updating a process response Jacobian matrix online, and calculating a multi-parameter cooperative adjustment amount; and updating the machine set value of the next period according to the adjustment amount. According to the method, the problems of control lag and parameter coupling are solved through prospective compensation and self-adaptive updating, and the control precision and robustness of multiple parameters in large-scale production of plastic products are improved.

The invention belongs to the technical field of multi-modal emotion recognition, and particularly relates to a multi-modal dialogue emotion recognition method based on uncertainty adaptive weighting. The method comprises the following steps: processing original modal information of each modal to obtain an alignment feature vector of each modal; calculating a cross-modal fusion feature of each cross-modal path based on the alignment feature vector of each modal; obtaining an uncertainty score of each cross-modal path according to the cross-modal fusion features of each path; obtaining the optimized weight of each path based on the initial weight and the uncertainty score of each path; and obtaining a predicted emotion category based on the optimized weights of all the paths and the cross-modal fusion features. According to the method, the expression ability of the non-text mode in the key emotion scene is remarkably enhanced while the semantic advantages of the text are kept.

The invention discloses a video content retrieval method and device based on television system voice interaction and a terminal, and relates to the technical field of video processing, and the method comprises the steps: when a video is played for the first time, extracting a picture frame from the video at a preset frequency, converting the picture frame into a multi-dimensional image feature vector and a corresponding video timestamp, and carrying out hierarchical storage in a database, constructing vectorized data containing visual semantic information; obtaining a voice retrieval instruction, performing intention recognition and semantic understanding, extracting a detection keyword, and generating a multi-dimensional retrieval feature vector; calculating a matching degree between the multi-dimensional retrieval feature vector and a multi-dimensional image feature vector of a video picture frame stored in a database, and screening out picture frames of which the similarity is higher than a preset similarity threshold to form a retrieval candidate matching set; and determining matched picture playing. The video content retrieval method is efficient, accurate and high in interactivity, and retrieval experience and operation efficiency of the user in the video watching process are remarkably improved.