218 results about "Shortterm Memory" patented technology

The invention discloses a truss structure wind-induced dynamic response prediction method and system based on physical enhancement. The method comprises the following steps: carrying out feature extraction and alignment fusion on input data containing condition parameters and wind speed time sequence data by utilizing a long short-term memory network and a physical enhancement attention mechanism; extracting multi-scale features from the fusion features through expansion convolution, and performing weighted aggregation on the multi-scale features; the physical priori knowledge of structural vibration is fused into position coding and a self-attention mechanism so as to carry out response prediction; and integrating physical model information of the truss structure and a dynamic control equation into a loss function, and calculating physical information residual loss so as to improve the physical interpretability of a prediction result. According to the method, data heterogeneity can be eliminated, complementary information can be fused, the multi-scale characteristic of wind-induced response is coped with, the accuracy and efficiency of wind-induced dynamic response prediction of the truss structure are effectively improved, and the physical interpretability and generalization ability are enhanced.

The invention specifically discloses a brain-like reinforcement learning method and system based on hierarchical experience playback, and relates to the technical field of reinforcement learning and brain-like computing. The method comprises the steps that S1, observation data are collected and preprocessed; s2, initializing an experience buffer pool, an actor network, a commentator network and a corresponding target network, and performing parameter initialization; s3, exploring noise is initialized, actions are selected from the actor network according to the current state and executed, and obtained experience samples are stored in an experience buffer pool; s4, obtaining a new sample from the experience buffer pool, and updating the short-term memory pool; s5, determining whether partial experience in the short-term memory experience pool is transferred to the long-term memory experience pool or not by using an attention discrimination module; and S6, updating parameters of the actor network, the reviewer network and the corresponding target network. By adopting the method, the experience utilization rate of the intelligent agent is improved, the reinforcement learning performance is improved, and the method has wide application potential in multiple fields.

The embodiment of the invention provides a context information management method and device based on hierarchical memory, equipment and a medium, and the method comprises the steps: analyzing a task request to generate a sub-task node sequence containing a target description and acceptance function, and arranging a plurality of agents to cooperatively execute a task, and after the audit is passed, a complete execution context is stored in a long-term memory, and a structured abstract is generated and stored in a short-term memory, so that the problem that the context management mechanism of the existing multi-agent system is single is effectively solved, stable maintenance of cross-session cognitive continuity is realized, and the efficiency of the multi-agent system is improved. The method improves the execution reliability of a complex long-process task, optimizes the information storage and retrieval efficiency through hierarchical memory, reduces the risk of semantic deviation accumulation and amplification, clarifies the task target and acceptance standard of each stage, enhances the multi-agent cooperation consistency, and improves the efficiency of task execution. And the execution efficiency and the task completion quality of the intelligent agent system in a complex application scene are comprehensively improved.

The embodiment of the invention provides a context construction method and device for an agent and a storage medium, and the method comprises the steps: obtaining the current input information of the agent in one interaction round, and carrying out the analysis to generate a current semantic representation; determining a semantic association degree between the current semantic representation and the context of the current dialogue task, and calculating a long-term value score of the current semantic representation; under the condition that the semantic association degree is greater than a first preset threshold value, storing the current semantic representation into a short-term memory library; under the condition that the long-term value score is greater than a second preset threshold value, storing the current semantic representation into a long-term memory library; when the intelligent agent needs to generate a response for the current input information, searching target memory content related to the semantics of the current input information from the short-term memory library and the long-term memory library; and combining the target memory content with the current input information to form prompt information, and inputting the prompt information into a large language model to generate a response for the current input information.

The invention belongs to the technical field of artificial intelligence, and discloses a large model memory management and optimization system and method based on a vector data lake. Comprising a multi-dimensional memory vectorization coding and hierarchical management module, an intelligent retrieval and dynamic updating module, and a memory optimization strategy and interactive evaluation and cyclic adjustment optimization module based on a reflection and backtracking mechanism. According to the large model memory management and optimization system and method based on the vector data lake, a multi-dimensional and hierarchical memory storage and dynamic retrieval framework is constructed, collaborative management of long-term / short-term memory is supported, efficient reasoning, knowledge reuse and continuous learning ability of an intelligent agent in a complex dynamic task are achieved, and the system and the method have the advantages that the system and the method are simple and convenient to operate. Therefore, the robustness, the adaptability and the task migration capability of the intelligent system are remarkably improved.

The invention discloses a data cleaning scheme generation method and device and a readable storage medium, and the method comprises the steps: carrying out the grouping of pre-training task data through employing a clustering mode, and obtaining a plurality of data groups; extracting data characteristics of each data packet based on a large model; generating a data cleaning scheme for the corresponding data groups by using the pre-training model in combination with data characteristics, and executing the data cleaning scheme; calculating a data cleaning index, and storing the data characteristics, the data cleaning scheme and the data cleaning index as short-term memory into a short-term memory module; evaluating the overall cleaning quality evaluation score of the data cleaning scheme by using a reflection module to obtain a reflection result; the data characteristics, the data cleaning scheme, the overall cleaning quality evaluation score, whether data cleaning succeeds or not and the reflection result serve as long-term memory and are stored in a long-term memory module, the data cleaning process can have traceability, and the data cleaning scheme adjusting efficiency is improved.

The invention discloses an enterprise-level agent system and a data processing method. The system mainly comprises a sensing module, a memory module, a planning module, a Routine agent planning engine, an arrangement engine and an execution module. According to the enterprise-level agent system provided by the embodiment of the invention, short-term memory, long-term memory and enterprise-level memory are combined through a complete closed-loop architecture of the sensing module, the memory module, the planning module and the execution module, and precipitation of personalized and organization-level knowledge is supported; and meanwhile, in combination with a standard business process template provided for an enterprise by a Routine agent planning engine and splitting of a task process by an orchestration engine, role agents are allocated for each sub-task, so that coordination and cross-system orchestration of complex tasks are realized, the construction threshold of intelligent applications is reduced, and the business efficiency is improved.

The invention discloses a hierarchical visual language navigation memory enhancement system in a cross-floor scene, relates to the technical field of intelligent navigation, and adopts the technical scheme that the hierarchical visual language navigation memory enhancement system comprises three core parts, namely a hierarchical visual semantic model construction module, a memory enhancement algorithm module and a navigation system integration module. The hierarchical visual semantic model construction module is used for constructing a basic visual feature layer, a floor semantic layer and a cross-floor semantic association layer; the memory enhancement algorithm module comprises a short-term memory module, a long-term memory module and a memory fusion and update strategy module; and the navigation system integration module is used for integrating the hierarchical visual semantic model and a memory enhancement algorithm into the system. Three-dimensional space characterization misalignment and path planning error accumulation can be effectively avoided, the accuracy, environment adaptability and decision-making efficiency of cross-floor navigation of the intelligent agent are remarkably improved, and the method has important technical innovation value and application prospects.

The invention belongs to the technical field of satellite remote sensing application, and discloses an active and passive satellite remote sensing fused ocean three-dimensional chlorophyll field reconstruction method and system. According to the method, satellite-borne laser radar data is subjected to correction, signal processing and biological optical model inversion, and a chlorophyll a concentration vertical section along an orbit is obtained; forming a training set by the passive satellite observation optical variable and the marine environment variable of which the profile is matched with the space-time, so as to train a long short-term memory (LSTM) neural network model; and utilizing the trained model to reconstruct a three-dimensional chlorophyll a concentration field of a target area according to passive observation and environment variables of the target area. By fusing the advantages of a satellite-borne laser radar ICESat-2 satellite and a passive optical remote sensing satellite, three-dimensional chlorophyll a concentration field detection based on active and passive fusion remote sensing is developed, the structure and function of a marine ecosystem can be deeply known, and three-dimensional dynamic observation of the ocean is realized.

The invention discloses a dialogue data storage method based on a big language model and a related device, and relates to the field of data storage, and the method comprises the steps: calling a memory encoder to obtain dialogue data of a user and an agent, employing the big language model to carry out key information extraction operation on the dialogue data, and obtaining key information, carrying out integrity verification and rationality verification on the key information; if the verification is passed, carrying out importance evaluation on the key information to obtain an importance evaluation value; and respectively storing the key information in a long-term memory database, a short-term memory database or a working memory database according to the importance evaluation value. The purpose of reliable storage of the dialogue data is achieved based on the large language model.

The invention discloses a robot emotion understanding and responding method, which comprises the following steps of: according to obtained multi-source data, setting a fusion initial weight through data integrity, and adjusting the fusion initial weight according to data confidence and environment quality to obtain a fusion weight so as to perform cross-modal feature fusion; according to the fusion feature, mapping to a two-dimensional emotion space, selecting a point with the maximum activation degree, and generating an emotion vector; based on the emotional state sequence, a response strategy vector is generated through a Linear Attention time sequence strategy network; based on a unified condition generation network, fusing the style vector and the response strategy vector to obtain an executable multi-mode instruction; calling a hierarchical memory network to obtain a short-term memory slot and a long-term sequence modeling unit which are used for modulating an executable multi-modal instruction; wherein the hierarchical memory network is trained and optimized through a loss function, and the loss function at least comprises style consistency loss, content consistency loss and memory retention loss.

The invention discloses a photovoltaic power prediction method based on improved empirical mode decomposition and an optimized long short-term memory network, and the method comprises the steps: firstly carrying out the preprocessing of abnormal value elimination, missing value filling, normalization and the like of photovoltaic power and related meteorological data, and improving the data quality; then, an improved empirical mode decomposition (EE-ANEMD) algorithm is adopted to decompose the preprocessed power sequence into a multi-scale intrinsic mode function component and a residual term, and high-frequency noise, intermediate-frequency fluctuation and a low-frequency trend are effectively separated; global optimization is carried out on the hidden layer unit number, the initial learning rate and the maximum number of training times of the LSTM network through an improved sparrow search algorithm (ISSA), finally, the optimized LSTM is utilized to carry out training prediction on each component, and results are fused and subjected to reverse normalization to obtain a final value. Experiments show that the test set RMSE of the method is reduced compared with that of a single LSTM, the mid-term prediction precision is remarkably improved, and reliable technical support is provided for power system dispatching, new energy consumption planning and photovoltaic power station operation and maintenance.

The invention provides a physical information long-short-term memory network site seismic response prediction method based on kinematics derivative constraint and data folding. The method comprises the following steps: collecting and preprocessing bedrock and ground surface seismic oscillation data; displacement, speed and acceleration supervision labels are constructed through surface acceleration integration; constructing a long-short-term memory network containing a data folding and unfolding module; physical constraint loss is constructed by utilizing kinematics derivative relations between displacement and speed and between speed and acceleration, and a network is trained by combining an adaptive weight strategy; and outputting earth surface displacement, speed and acceleration time history by adopting the trained network. According to the method, under the condition that constitutive parameters of a soil layer are not explicitly given, physical consistency constraint can be applied to the field seismic response prediction process, the long-time program train training burden is reduced through sequence folding processing, and the method is suitable for field seismic response prediction.

The invention relates to the technical field of coal mine rock burst monitoring and early warning, in particular to a rock burst mine micro-seismic data b value calculation method and device, electronic equipment and a computer readable storage medium, and the method comprises the steps: obtaining original micro-seismic time sequence data of a target monitoring area; preprocessing the original micro-seismic time sequence data to obtain preprocessed micro-seismic time sequence data; performing normalization processing on the preprocessed micro-seismic time sequence data to obtain standardized micro-seismic time sequence data; performing wavelet multi-scale decomposition to obtain a low-frequency approximate component and a high-frequency detail component; carrying out linear fitting by utilizing a Guyden-Ricker law to obtain a multi-scale b-value data set; and constructing and training a long-short-term memory network model, and obtaining a multi-scale b value fusion result as a final b value. Through a wavelet multi-scale decomposition technology, micro-seismic data is adaptively decomposed into components reflecting different time scales, and the limitation of manually setting a single sliding window is eliminated.

The invention discloses a multi-agent-based cross-station intelligent inspection and operation and maintenance method and system, and the system comprises an inspection expert agent, a tool use agent, a code generation agent, an overhaul expert agent and a memory module. Calling a tool using agent to obtain corresponding power station information generation suggestions, and issuing corresponding tasks after user confirmation; a code generation agent receives the task, a tool use agent is called to obtain a corresponding power station model version and hardware equipment information generation script, and script parameters are dynamically adjusted; and the maintenance expert agent obtains the script operation detection model, calls the tool use agent to obtain corresponding power station monitoring information, detects the monitoring information by using the detection model, stores the detection result to the short-term memory module, and gives a maintenance suggestion and adjusts a detection strategy by combining the detection result with the maintenance knowledge information in the long-term memory module. According to the invention, efficient, intelligent and sustainable automatic task management and execution are realized.

The invention relates to a monitoring and diagnosing method, device and equipment of a power transformer and a medium. The method comprises the following steps: acquiring multi-dimensional monitoring data acquired by a sensor array in a full state of a transformer, and denoising to obtain a clean data sequence; inputting the cleaning data into a long short-term memory network to extract oil deposition time sequence dynamic characteristics, and combining a trend quantification model to study and judge a deposition accumulation trend and generate a trend curve; if the accumulation amount or the rate in the curve exceeds a preset threshold value calibrated based on historical fault data and industry standards, multi-source deposition risk key influence factors are collected, and after preprocessing, an improved weighted evidence theory is adopted for fusion to obtain deposition risk dominant factors; and inputting the factor into a convolutional neural network to locate a potential fault type, generating an adaptive diagnosis report, and if a high-risk fault exists, triggering and outputting a real-time intervention instruction. According to the method, early recognition, accurate diagnosis and timely intervention of oil deposition associated faults are realized, and the operation safety and stability of a power system are improved.

The invention provides a semiconductor silicon single crystal growth value prediction method based on event trigger learning, and belongs to the technical field of silicon single crystal growth prediction. Comprising the following steps: constructing a value state space model according to a silicon single crystal growth mechanism; embedding the value state space model into a long short-term memory network to construct a network model; an attention mechanism is introduced into the network model, weighted fusion processing and physical constraint mapping processing are carried out in sequence, and the network model is constructed; based on an event triggering learning mechanism, training and adjusting the network model by utilizing historical time sequence data of the semiconductor silicon single crystal growth value to obtain a trained network model; and inputting the technological parameters at the current moment into the trained network model, and predicting the growth value of the semiconductor silicon single crystal at the next moment. According to the method, the accuracy, the real-time performance and the self-adaptive updating capability of semiconductor silicon single crystal growth value prediction can be improved.

A computer-implemented method and an apparatus for automatic modulation recognition that enables the detection and identification of modulation schemes in received raw signals with a signal receiving unit (70) without prior information about the raw signal detail, characterized by; comprising at least one computing unit (10) configured to transforming received raw signals from time domain to frequency domain including the noise in the signal with segmenting the signal and computing its modulation in multiple image with the spectrogram extraction process in order to capture temporal dependencies and sequential information by treating the raw signals received from signal receiving unit (70) as images; augmentation of the data for increasing the dataset size for increasing the accuracy with the limited data; training the data for enabling the network to learn spatiotemporal relationships based on a predefined or a precalculated signal-to-noise ratio (SNR) level; applying the data to one algorithm of two, which are convolutional neural network (CNN) and convolutional neural network (CNN) long short-term memory network (LSTM) hybrid algorithm.

The invention relates to the technical field of electric connector weak light assembly vibration compensation alignment methods, particularly discloses an electric connector weak light assembly vibration compensation alignment method, and aims to solve the problems of low electric connector assembly positioning precision and high pseudo soldering rate caused by vibration in a weak light environment. The method comprises the following steps: monitoring the thermal radiation change of a vibration source in real time through an infrared thermal imaging sensor, and pre-judging a vibration acceleration vector in combination with a lightweight long-short-term memory neural network; synchronously triggering dual polarization angle imaging to respectively capture a housing reference mark and an enhanced terminal metal feature; the vibration data and the polarization image flow are fused in the field programmable gate array, a self-adaptive optical flow algorithm is adopted to resolve the displacement compensation amount, and a piezoelectric ceramic actuator is driven to carry out submicron order reverse displacement compensation; when the compensation amount exceeds a threshold value, a low-delay correction instruction is sent to the robot through an Ethernet control automation technology bus, and the steady-state error is ensured to be smaller than 0.3 micrometer by combining a strain gauge fine tuning feedback loop. According to the technical scheme, the assembling precision, efficiency and production line adaptability are remarkably improved.

The application relates to the technical fields of artificial intelligence and man-machine natural language dialogue, and provides a general memory management method and system based on a language model, which extracts memory information needing to be memorized from dialogue content with the language model, generates temporary memory, integrates the temporary memory to obtain persistent memory about a dialogue user, and is integrated into a dialogue process of the dialogue user and the language model to form memory of the dialogue user, so that the dialogue of the dialogue system is controlled, memory type classification management is performed according to the importance of the memory information in the persistent memory, so that the persistent memory contains short-term memory and long-term memory, the importance of the memory information in the persistent memory is updated and managed, and the short-term memory and the long-term memory are forgotten and upgraded, so that flexible and efficient storage and calling of user memory independent of the model type are realized, and the humanized, continuous and personalized dialogue experience in the man-machine dialogue process is improved.

The invention belongs to the technical field of software development, and particularly provides a research and development-oriented long and short-term memory framework construction method and system, which adopts a layered memory architecture to construct four core modules including a short-term memory compressor, a medium-term memory aggregator, a long-term memory graph and a cross-layer memory router. The system takes multi-source input such as research and development dialogues, code snippets and project documents as a starting point, extracts research and development elements through semantic analysis and entity recognition technologies, compresses lengthy dialogues into structured short-term memory by utilizing an attention distillation mechanism, upgrades high-frequency short-term memory into medium-term knowledge fragments based on a time sequence attenuation algorithm, and improves the research and development efficiency. And constructing a long-term knowledge graph containing developer portraits, project dependence and normative standards by adopting a graph convolutional network. Context understanding accuracy, multi-round dialogue continuity and personalized service quality of a large model in a research and development scene are remarkably improved, and the method is suitable for mainstream research and development tool scenes such as IDE plug-ins, code review and architecture design.

The invention discloses an interactive response method, device and equipment for multi-modal sensing data and a storage medium, and relates to the technical field of intelligent interaction.The interactive response method for the multi-modal sensing data comprises the steps that the multi-modal sensing data is obtained, and a user mental evaluation result is constructed based on the multi-modal sensing data; identifying user identity information according to the multi-mode sensing data, and determining short-term memory information and long-term memory information according to the user identity information; when a demand instruction of a user is received, performing retrieval in the short-term memory information and the long-term memory information according to the demand instruction to obtain a retrieval result; and generating interaction response content according to the retrieval result and the user mental evaluation result, and performing interaction. Through multi-modal perception and dynamic memory management, long-term continuous personalized interaction is realized, and the user interaction experience is improved.

The application discloses an advertisement optimization method, system, device and medium. The method comprises the following steps: obtaining an error value between a current advertisement effect and an expected effect; calculating a nonlinear PID control signal by using a nonlinear PID algorithm according to the error value; obtaining historical data of the effect of the advertisement, and predicting the error by using a long short-term memory network on the historical data to obtain error prediction data; converting the error prediction data by using a nonlinear function to obtain a feedforward compensation signal; and optimizing the advertisement according to the nonlinear PID control signal and the feedforward compensation signal. Compared with the related art, the application optimizes the PID algorithm by introducing a nonlinear function, and improves the response speed of the advertisement optimization system while improving the accuracy of the advertisement optimization system by combining the long short-term memory network (LSTM) for feedforward compensation prediction.

The invention relates to the technical field of water treatment, discloses an intelligent flocculant adding method and system based on water quality dynamic data, and aims to solve the technical problems that an existing adding mode is extensive, lagged in control and difficult to adapt to dynamic water quality fluctuation. The method comprises the following steps: collecting turbidity and flow data of a water inlet end; carrying out feedforward prediction by utilizing a long-short-term memory network model, calculating a theoretical dosage and driving a metering pump to execute; and the effluent turbidity is monitored in real time, and feedback correction is performed by using a PID algorithm. The system comprises an online monitoring unit, an intelligent control unit, a precise adding execution unit and an effect feedback unit. Through composite control of feedforward prediction and feedback correction, advanced and accurate addition of the flocculant is realized, the drug consumption cost is reduced, the impact load is effectively dealt with, and the stability of the effluent quality and the intelligentization of system operation and maintenance are ensured.

The invention relates to the technical field of data processing, and discloses a session data processing method, device and equipment and a readable storage medium, the session data processing method comprises the following steps: determining a theme tag based on session data of a user; matching the topic tag with a keyword index in the long-term memory data, and screening out a candidate memory data set meeting a preset condition; sorting the candidate memory data set according to the keyword correlation and the time information to construct a short-term memory data set; historical memory content related to the current session data is retrieved in the short-term memory data set, and target session content is generated based on a retrieval result. Through a subject-driven classified storage and preloading mechanism, the memory retrieval efficiency is remarkably improved, the repeated access overhead of a long-term memory library is reduced, the response delay is effectively shortened, the context coherence and individuation ability of multiple rounds of dialogues are enhanced, and the user experience is comprehensively improved.

This invention discloses a lightweight driver distraction behavior detection method based on CNN and ViL, belonging to the field of driver distraction detection. It constructs a hybrid model combining a Convolutional Neural Network (CNN) and a Visual Long Short-Term Memory (ViL) network for driver distraction behavior detection. The hybrid model performs five stages of processing. In the first two stages, local feature extraction is mainly performed. The DEC module is used to extract and downsample local features, while the PDC module further enriches the extracted local features. In the last three stages, the MlgViL module is used to process and fuse local and global information. This invention efficiently fuses local and global information, preserving fine-grained local information while capturing global contextual information, thus improving the overall performance of the model and achieving high accuracy and real-time performance with limited computing resources.

The application discloses a grid-connected fault comprehensive evaluation strategy for a network-constructed hybrid energy storage power station. By adopting a hybrid evaluation framework that fuses an analytic hierarchy process, a convolutional neural network, a long short-term memory network and an attention mechanism, the evaluation precision of a complex fault scene is improved while the model interpretability is maintained, and accurate quantification of the grid-connected fault severity of the network-constructed hybrid energy storage power station is realized.

A method for predicting a next time step data element in a set of time series data includes receiving a one-dimensional time series data set and converting the time series data set to a two-dimensional time series data set. The two-dimensional time series data set is provided to an input of a convolutional neural network-long short term memory (CNN-LSTM) model. The CNN-LSTM model generates the next time series step prediction of the two-dimensional time series data. The method compares the prediction to an actual next time series step and responds to a difference exceeding a first threshold by incrementing an outlier counter. When the outlier counter exceeds a predefined size the method alters the CNN-LSTM model. In addition, a visualization of the next time series step prediction of the two-dimensional time series data is generated.

The invention relates to the technical field of power system stability control, and provides a new energy black-start partition optimization method considering unconventional risks, which comprises the following steps: simulating unconventional risks possibly encountered by a power system by adopting Monte Carlo sampling; extracting the output characteristics of wind power under the unconventional risk by adopting variational mode decomposition, and performing output prediction in combination with a long-short-term memory neural network; a power system recovery partition is divided based on an LPA algorithm, and the LPA algorithm is improved by considering a partition size balance principle, so that the partition area is more balanced, and the tag oscillation effect is reduced. According to the method, the black-start partition optimization model fusing new energy output uncertainty modeling and unconventional risk scene generation is designed, and the partition scheme is cooperatively solved by adopting the improved label propagation algorithm, so that the robustness and the recovery efficiency of the black-start partition are effectively improved; and rapid and stable power supply of a power system under complex disturbance can be realized.