147 results about "Data aided" patented technology

The invention provides a question and answer method and device based on a large language model, and the method comprises the steps: obtaining a current dialogue state which comprises a query text in a current dialogue; based on the query text, target memory chain data is retrieved from a memory chain database, the memory chain database comprises multiple pieces of memory chain data, and the multiple pieces of memory chain data are obtained based on knowledge extraction of historical dialogue data; obtaining a prompt text based on the query text and the target memory chain data; and taking the prompt text as the input of the large language model, and outputting an answer corresponding to the query text. According to the question and answer method based on the large language model, knowledge extraction is performed on the historical dialogue data to construct the memory chain data, the memory chain data assists the large language model to generate higher-quality answers, and the user experience is improved.

The invention relates to a hot topic chatting method, system and device based on an intelligent agent and a storage medium. The method comprises the steps of extracting key intention information through a large language model according to current input data and historical dialogue data; constructing a topic vector database and automatically updating; retrieval is performed based on cosine similarity, and hot topics are screened out; the hot topics are integrated into a Prompt template, an auxiliary tool is called in real time to obtain auxiliary information, and a large model is used for generating reply content based on the Prompt template according to the current input data and the auxiliary information; and storing the current input data and reply contents, and updating the historical dialogue data. Personalized emotional appeals of a user are analyzed according to data input by the user through a large language model, key intention information is extracted, a topic vector database is constructed, topics most related to current user input are screened out, and by means of auxiliary tools, the real-time performance of hot topics can be guaranteed, and emotional resonance with the user can be enhanced.

In order to solve the problem that the snow depth predicted by a sentinel No.1 data set is difficult to meet high spatial resolution and high temporal resolution at the same time, the invention provides a snow depth prediction method, a storage medium and electronic equipment, and the method comprises the steps: obtaining sentinel No.1 data, auxiliary data and station data about the snow depth, constructing a geographic time weighted snow depth parameter based on the site data; carrying out resampling and space-time matching on the data to obtain a first data set; inputting the first data set into an integrated deep learning model to obtain a preliminary snow depth; obtaining time sequence climate data and geographical time data of the research area, performing resampling, and converting the preliminary snow depth data into time sequence data; the resampled time sequence climate data and geographical time data and the time sequence data of the initial snow depth form a second data set; and inputting the second data set into the ResNet-BiLSTM time sequence model to obtain snow depth data. According to the invention, snow depth prediction data with high temporal-spatial resolution can be generated.

The invention provides a PC (Personal Computer) end quantitative model reasoning method. The method comprises the following steps: S1, loading input information; s2, deserializing the model; s3, operator calculation is carried out; and S4, storing the data, and selecting a dump data result. The quantitative model reasoning on the PC end is used for simply reasoning the model, verifying the result of the model and obtaining the data of the network layer of each stage reasoned in the model, and the PC end reasoning is used for correcting the data, assisting the quantitative model, reasoning the float model, post-quantitative model and training the quantitative model and verifying the correctness of the data. According to the method, model reasoning is convenient and effective, model data verification is facilitated, and data information of a quantitative model is provided.

The invention discloses a disaster comprehensive management method and system and a computer program. The method comprises the following steps: acquiring a historical natural information data set; constructing a training decision tree to identify disaster types, and obtaining a disaster type identification model; constructing and training a full-connection neural network to predict a disaster assessment score, and obtaining a disaster assessment score prediction model; natural information data are monitored and acquired in real time, standardization processing and feature extraction are performed, analysis is performed through the trained disaster type identification model and the disaster assessment score prediction model, and disaster types and disaster assessment scores are acquired; acquiring a disaster severity level; and processing by adopting a preset strategy. According to the technical scheme provided by the invention, a plurality of functions are integrated together, so that the management efficiency and the response speed are improved; real-time data acquisition and processing are realized, disaster risks can be found in time, and corresponding early warning and countermeasures can be taken; the artificial intelligence algorithm is used for analyzing disaster data, decision making is assisted, and the accuracy and efficiency of dealing with disasters are improved.

Certain aspects of the present disclosure provide techniques for method for wireless communications by a user equipment (UE) generally including receiving a plurality of code blocks (CBs) from a network in one or more orthogonal frequency division multiplexed (OFDM) symbols and decoding the CBs using an iterative process performed on each OFDM symbol, wherein the iterative process involves data aided channel estimation and each iteration of the process involves demodulation and decoding.

The embodiment of the invention provides a sensing signal determination method, a storage medium and an electronic device. The method comprises: determining time-frequency domain resources and data transmission parameters according to the channel quality of a channel, the communication demand of a terminal and the sensing demand of a base station, the time-frequency domain resources comprising time domain resources and frequency domain resources; sending the allocation indication information of the time-frequency domain resources and the data transmission parameters to the terminal, and obtaining data signals sent by the terminal at the time-frequency domain resource position of the channel according to the data transmission parameters; and reconstructing the data signal to obtain a sensing signal of the time-frequency domain resource position where the data signal is located, and sensing the channel and the environment based on the sensing signal. The method at least solves the problem that in the related technology, communication data assisted perception communication and sensing fusion cannot be achieved.

The invention discloses a remote auxiliary diagnosis method and system based on multi-modal medical data, and relates to the technical field of auxiliary data processing, and the method comprises the steps: obtaining a customized remote auxiliary diagnosis task containing sample multi-modal medical data and corresponding label data, and then carrying out the remote auxiliary diagnosis according to the customized remote auxiliary diagnosis task, the target deep learning model is optimized by adopting a multi-strategy hybrid optimization algorithm, the optimized target deep learning model is deployed on a cloud server, and finally, the deployed target deep learning model is called to recognize the to-be-recognized multi-modal medical data uploaded by the medical staff to obtain a data auxiliary processing result. According to the invention, the method can achieve the custom remote auxiliary diagnosis task of a medical worker, assists the medical worker in improving the processing efficiency of multi-modal medical data, employs a multi-strategy hybrid optimization algorithm to optimize the target deep learning model, and can improve the accuracy of processing the custom remote auxiliary diagnosis task.

The invention discloses a distribution network project cost method and system for auxiliary auditing of spatial metadata, and relates to the technical field of distribution network project cost auditing, and the method comprises the steps: integrating various data, constructing a database, and evaluating the reasonability of a line; judging rationality of project amount, site layout and power distribution network project cost; and constructing an intelligent auditing model to assist intelligent review. According to the distribution network project cost method for auxiliary auditing of the spatial metadata, through association and model construction of the spatial metadata and the project cost data, the project cost can be rapidly and preliminarily evaluated and screened, the auditing efficiency is improved, the project cost is comprehensively analyzed from the spatial dimension, and the auditing efficiency is improved. According to the method, the probability of human negligence and errors is reduced, the accuracy of the auditing result is improved, the auditing method based on the spatial metadata is beneficial to discovering problems related to unreasonable spatial layout in the project cost, the project benefits are improved, and the method achieves better effects in the aspects of auditing efficiency, accuracy and project benefits.

The invention provides an atlas updating method and device based on a large model, electronic equipment and a medium, information is extracted from large-scale text data by utilizing a natural language processing model, and the method mainly comprises the steps of data collection and preprocessing, entity extraction, relation extraction, atlas updating, model evaluation and iteration and the like. According to the method, an automatic crawler technology, multi-language text processing, transfer learning, visual data auxiliary entity extraction, relation extraction based on a pre-training language model and other technologies are comprehensively applied, and multiple beneficial effects of graph timeliness, multi-language adaptability, model generalization ability improvement, extraction accuracy improvement and the like are achieved. The atlas updating and continuous optimization process is optimized by means of an incremental updating strategy, a dynamic learning system, user feedback data, a conflict resolution strategy and the like. In conclusion, according to the method, the accuracy, the integrity and the user friendliness of the atlas are remarkably improved, meanwhile, the dependence on the annotation data is reduced, and the overall calculation efficiency and the model generalization ability are improved.

The invention relates to the technical field of government affair data processing and intelligent decision making, and particularly discloses an index management-based data aided analysis, prediction, research and judgment system, which is characterized in that a multi-level index management system is constructed, an association relationship between indexes is defined, and a dynamic update threshold value is configured; the method comprises the steps of collecting original operation data from multiple government affair data systems, and generating a standardized data set through quality verification and content integration; establishing a multi-index cooperative monitoring network by performing correlation analysis and abnormal feature recognition on the data set; deducing a plurality of development paths in a future time period based on state transition analysis of a historical sequence, and generating a prediction result including trend pre-judgment and risk level; and finally, constructing a decision scheme tree, generating a visual analysis report through multi-dimensional scoring, continuously optimizing parameters based on implementation feedback, and forming a closed loop of analysis prediction and decision support. According to the invention, deep fusion and intelligent research and judgment of cross-domain government affair data are realized.

A compute server of a distributed cloud computing network receives an inference request. The inference request triggers execution of code that is related to an AI application that interacts with the inference request and causes its input to be run through an AI model. Helper code executing on the compute server transmits an inference request to the AI model, where this inference request includes a prompt and one or more function definitions, and where the AI model is executing in the same datacenter as the compute server. The helper code receives a result of the AI model executing the inference request, where it includes structured data for calling a function. The helper code calls the function using the structured data. The helper code receives a response to the called function. The code processes a response to the inference request based at least on the received response to the called function.

Methods, device and data stream are provided to encode unaligned multi-view plus depth images associated with assistance metadata for generating depth maps aligned with the unaligned color views. At the encoding stage, for an unaligned color view, the contribution of each depth map is evaluated, and the objects represented are mapped. Picture patches are cut to be homogeneous in terms of objects and contributing views. Metadata describing this homogeneous cut are encoded together with the atlas. At the decoding stage, to generate a depth map aligned with the unaligned color view, only the depth maps referenced in the assistance metadata are warped.

The invention discloses a sea wave height inversion algorithm based on C-band SAR, and relates to the field of sea wave height inversion, and the algorithm comprises the steps: selecting mode data in offshore satellite data; eRA5 wind field data and wave height data are selected, mode data are processed after time matching, SAR image sub-images are selected, and then space matching is carried out on the SAR image sub-images and the ERA5 wind field data and the wave height data; obtaining training data by using the SAR image sub-image after space matching, the SAR data auxiliary information file, the ERA5 wind field data and the wave height data to construct a training data set; and constructing an inversion model, inputting the training data into the inversion model, and finally obtaining the SAR sea wave height. Ocean wave height inversion is carried out by adopting complete polarization data, the precision is high, the application threshold is high, and the application range is small; meanwhile, the problems of inversion precision and applicability are comprehensively considered, more SAR information is used for sea wave inversion, and meanwhile the application range of an inversion model is guaranteed.

An assisted driving system includes: inputs configured to receive sensor data, telematics data, and driver response data; an assisted driving module configured to, based on the sensor data and telematics data generate first non-urgent notification messages; and a ranking module configured to rank the first non-urgent notification messages based on the driver response data. The assisted driving module is configured, based on the ranking of the first non-urgent notification messages, to output via at least one output device the first non-urgent notification messages such that saliency of each of the first non-urgent notification messages is based on the respective ranking of that non-urgent notification message.

The present invention provides an OPC UA server search method based on port query. By constructing a virtual port in a user terminal, the method obtains the IP address information of all OPC UA servers in a local area network, thereby facilitating a rapid search for a matching OPC UA server, thereby achieving a fast connection between the user terminal and the OPC UA server, and improving the reliability and efficiency of the data-assisted computing processing provided by the OPC UA server to the user terminal.

The application belongs to the field of OFDM channel estimation, and relates to a data-assisted short-wave OFDM sparse channel Turbo iteration estimation method, which comprises the following steps: S1, receiving a signal transmitted through a channel and processing the signal to obtain a receiving end symbol matrix Y; S2, performing channel estimation and equalization on Y to obtain an estimated channel and an OFDM signal pair, and performing Turbo iteration on the estimated channel and the OFDM signal pair to obtain a symbol variance; S3, selecting a virtual pilot in the signal according to the symbol variance; S4, performing channel estimation and equalization on Y according to the virtual pilot to obtain a current estimated channel and an OFDM signal; S5, performing Turbo iteration on the signal to obtain a current symbol variance; and S6, if a preset maximum iteration number is reached, obtaining a final estimated channel; otherwise, returning to step S3 to select a virtual pilot in the signal according to the symbol variance; and the application selects a virtual pilot in the signal according to the symbol variance, and uses the virtual pilot to perform estimation again, so that the spectral utilization rate is improved and the channel estimation performance is improved.

The invention discloses a medical data aided decision-making method and device fusing multi-modal data, and relates to the technical field of artificial intelligence, and the method comprises the steps: obtaining the original medical data of a to-be-predicted patient; aligning semantics among the data of various modalities, and generating personalized feature vectors of the patient to be predicted; determining an activation weight according to the feature information of the personalized feature vector; according to knowledge data in the original medical data, a clinical rule is converted into a fine-adjustable constraint condition, and according to the constraint condition and the activated identification expert layer, feature analysis is performed on the personalized feature vector to obtain an illness state analysis state vector; assembling cue words, and reasoning the cue words; and generating a plurality of candidate reasoning paths, determining an optimal reasoning path in the candidate reasoning paths, and generating an optimal reasoning result by using the optimal reasoning path. The invention provides an LLM for performing dynamic verification and adjustment by fusing multi-dimensional contrast signals in real time in the process of medical data analysis and auxiliary decision generation.

The application provides an electromagnetic safety restoring force digital twin training method and system, constructs a test twin space, determines a test equipment group based on electromagnetic test data, and associates real interference data; determines interference dimensions based on equipment properties of each test equipment in the test equipment group, generates false interference data according to the interference dimensions; updates the state of the test equipment according to the real interference data, generates actual data, and directionally replaces the running data of the corresponding dimensions based on the false interference data to generate false data; collects calling information of the training end to the auxiliary plug-in, generates error area guide data according to the comparison information of the calling information, the actual data and the false data, and the auxiliary plug-in one-to-one corresponds to the running data of each dimension, so that the real interference data and the false interference data can be fused, the training complexity is improved, and the disposal efficiency of an operation and maintenance personnel for an electromagnetic safety event is improved.

The invention discloses an over-the-horizon railway safety management method and system. The over-the-horizon railway safety management method comprises the following steps: collecting video data of each key road section through a ground video monitoring subsystem; when there is no line node, the vehicle-mounted terminal carries out sampling and calculates the distance between the corresponding line node and the key road section; when the line nodes are known, the vehicle-mounted terminal obtains the real-time coordinates of the train, and the line node closest to the real-time coordinates serves as the current position of the train; when the distance between the current position of the train and the starting point of the nearest key road section to be driven in is smaller than or equal to the preset approaching distance, image data stream transmission between the ground video monitoring subsystem and the vehicle-mounted terminal is started through the over-the-horizon train-ground safety monitoring platform; and when the terminal distance between the current position and the nearest driven-out key road section is greater than or equal to a preset far distance, stopping image data stream transmission. According to the invention, high-definition image data aided driving can be realized in the high-speed running process of the train.

The invention relates to the technical field of intelligent traffic, in particular to a lane positioning system and method for vehicle-road collaborative awareness, which comprises a data collection module, a data transmission module, a storage library module, a central processing module and an instruction output module, and is characterized in that the data collection module is in control connection with the data transmission module, and the data transmission module is in control connection with a preprocessing module; the automatic driving module assists the vehicle in automatic driving according to the positioned data, the system and the automatic driving function are linked, and convenience and reliability are achieved; the system runs safely and stably in the system normal monitoring module, and an offset correction system is used in the lane offset correction module to correct lane offset, so that lane offset early warning and automatic correction are realized, and potential safety hazards are eliminated; after offset correction in the repeated positioning module, confirmation is carried out again, and result elimination early warning is fed back in the system feedback module; the system locates a lane by means of a sensor and a camera, can be used in a place which cannot be covered by a positioning system, and has very high environmental adaptability.

The embodiment of the invention discloses an elevator fault monitoring method and device based on running speed, a medium and equipment, relates to the technical field of intelligent monitoring, and aims to solve the problem of low effectiveness of elevator fault monitoring based on running speed data in the prior art. According to the method, multiple pieces of operation speed data are obtained through different speed monitoring algorithms, one piece of data is determined as target data and used for elevator fault monitoring, under the condition that the target data are missing, other data is used for assisting in complementing the target data, homologous data is not used for complementing, the influence caused by faults of collection equipment is avoided, and the accuracy of elevator fault monitoring is improved. The completeness and effectiveness of the data are improved, and then the complemented operation speed data are used for monitoring, so that the effectiveness of fault monitoring can be improved.

The utility model provides a pulse resonance deicing device, including solar energy device, master control box and external excitation source device, solar energy device includes solar panel and fixed steel support, solar panel is fixed on the power tower through fixed steel support, the master control box is equipped with communication board, master control board, power panel and storage battery inside, and the external excitation source device is equipped with the communication board, master control board, power panel and storage battery inside. The main control board is in signal connection with the far-end monitoring platform through the communication board, the solar panel is connected with the power panel, the power panel is connected with the storage battery, the storage battery is connected with the main control board, and the external excitation source device comprises a pulse generation module, a resonant frequency detection module and a resonance generation module. The pulse generation module, the resonant frequency detection module and the resonance generation module are connected with the power panel and the main control panel respectively. The device is convenient to install and high in maintenance-free degree. While high-efficiency, low-consumption, pollution-free and safe deicing is realized, the ice and snow coverage condition of the current power transmission line can be remotely transmitted to a remote monitoring system, and related data is provided to assist a user in decision making.

The invention relates to the technical field of servers, and discloses a node assembly and a server system.The node assembly comprises a node component and an auxiliary component, the node component comprises a first shell and a computing node arranged in the first shell, and the computing node is used for processing and storing data. The auxiliary component comprises a second shell and a transmission panel arranged in the second shell, and the transmission panel is used for providing an external auxiliary interface for the computing node so as to assist the computing node in working. In the first direction, the second shell is detachably connected to one end of the first shell, the first shell and the second shell can be fixed through a first installation position or a second installation position, the first installation position and the second installation position are arranged in the second direction, and the first direction intersects with the second direction.

The invention relates to the field of laser communication, and provides a parallel low-complexity frequency offset compensation method, device, equipment and medium, and the method comprises the steps: calculating the M-power variance of N M-PSK signals which are input in parallel, and carrying out the first frequency offset compensation of an original signal through the difference result of N points at intervals; the M-PSK signal is a PSK signal of which the modulation order is M; and solving an N-point M-power spectrum for the signal after the first frequency offset compensation, and performing second frequency offset compensation for the signal after the first frequency offset compensation by using the spectral line position of the M-power spectrum. The two-stage frequency offset compensation method provided by the invention has the characteristics of high dynamic range, low resource consumption, no need of data assistance and support of burst signal processing, and is particularly suitable for a frequency offset compensation scene in a high-speed laser communication link.

The invention discloses a method for realizing MODBUS communication based on a serial port, a slave station terminal and equipment, and the method comprises the steps: firstly initializing the serial port, starting to monitor a communication state after an auxiliary state machine is switched to a state of waiting for receiving data, and waiting for receiving frame data when a first state condition is satisfied; if the first character data is received, the auxiliary state machine is switched to a data receiving state, and receives frame data when monitoring that the communication state meets a second state condition; if the serial port idle interruption is triggered when the frame data is received, the auxiliary state machine is switched to a data processing state, and the received data is processed according to the MODBUS protocol when it is monitored that the communication state meets a third state condition; and finally, when the processing result data needs to be returned, the auxiliary state machine is switched to a data returning state, and the processing result data is sent to the master station terminal under the condition that the communication state meets a fourth state condition. The communication state is monitored in real time through the auxiliary state machine, so that the communication error-tolerant rate and stability are improved.

The present invention discloses a calculation system and method for the thermal value of the driving road of a new energy vehicle in the field of vehicle networking technology, including: obtaining the position data of the new energy vehicle; converting the position data through reverse geocoding to obtain a list of roads closest to the new energy vehicle; calculating based on the distance from each road in the road list to the position of the new energy vehicle, storing the distance score from each road to the position of the new energy vehicle after calculation; taking the sum of the distance scores of each road within a specified period as the thermal value of the driving road of the new energy vehicle. The present invention combines the characteristics of new energy vehicles, introduces technical means such as big data and geographic information systems, has high calculation efficiency, and has a certain degree of generality. By calculating the thermal value of the driving road of new energy vehicles, it provides certain data assistance and reference for enterprises to build charging stations and maintenance stations.

The invention relates to an electroencephalogram signal preprocessing method and a related device thereof.The method comprises the steps that initial electroencephalogram signals collected by a multi-channel electrode array and at least one kind of auxiliary data collected synchronously with the initial electroencephalogram signals are obtained, and the auxiliary data comprise electrode impedance values and / or physiological state data; wherein the resistance reactance value indicates an electrode impedance value corresponding to the acquisition channel associated with the initial electroencephalogram signal, and the physiological data at least comprises motion posture data and / or physiological index data; performing multiple filtering processing on the initial electroencephalogram signal to obtain a filtered electroencephalogram signal; inputting the filtered electroencephalogram signals and at least one kind of auxiliary data into a signal restoration model, and optimizing the candidate electroencephalogram signals to obtain optimized target electroencephalogram signals. According to the scheme, the electroencephalogram signals can be preprocessed by dynamically adjusting the filtering parameters and utilizing the impedance value and the auxiliary data, and the quality of the electroencephalogram signals is improved.

The utility model relates to a 3D curved surface data acquisition device. The device comprises a first linear motion platform; the supporting seat is arranged on the first linear motion platform in a sliding manner; the rotating platform is rotationally arranged on the supporting seat, a product with a 3D curved surface is arranged on the rotating platform, and the first linear motion platform drives the product to move to a detection position; the motion extension direction of the second linear motion platform is different from that of the first linear motion platform; the sliding support is arranged on the second linear motion platform in a sliding mode, and the laser is obliquely arranged on the sliding support; and the control unit is used for controlling the first linear motion platform, the rotating platform and the second linear motion platform to move, so that the laser obtains 3D curved surface data of the product. According to the utility model, the 3D curved surface product is automatically moved, the curved surface data is comprehensively acquired by laser, and the high-precision detection of the 3D curved surface is assisted.

The invention discloses a dynamic perception-based maintenance auxiliary equipment monitoring system and method, and relates to the technical field of auxiliary equipment monitoring, and the system comprises the steps: dividing a monitoring region according to a monitoring range, marking a generator set, collecting historical operation data, auxiliary tool data and fault type data, and building a fault database for the monitoring region; calling the fault information of the monitoring area, analyzing the demand condition of the fault type for the auxiliary tools based on historical data, predicting the demand condition of the auxiliary tools in a monitoring period, analyzing the storage condition of the auxiliary tools in the monitoring range, and prompting management personnel to supplement until the auxiliary tools in the monitoring range are sufficient. According to the method, the fault database is established, maintenance records of effective maintenance personnel are integrated, data interference of invalid personnel is eliminated, and the actual requirements of different fault types for various auxiliary tools are objectively judged.